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


Channel log from 28 January 2008   (all times are UTC)

00:09cyberorg has joined #ltsp
00:11F-GT has quit IRC
00:11OvaKill has joined #ltsp
00:14zodman has joined #ltsp
01:12misui has joined #ltsp
01:35armin2 has joined #ltsp
01:55mikkel_ has joined #ltsp
02:01arbu_ has joined #ltsp
02:03
<vagrantc>
warren: yeah, it was a pretty spontaneous decision ... just keep hearing gentoo folks expressing interest, and am relatively close to dberkholz
02:03armin2 has quit IRC
02:14arbu_ has quit IRC
02:16
<Q-FUNK>
vagrantc: we just hired a Gentoo guy, who might get around it.
02:20
<dberkholz>
Q-FUNK: who?
02:20
<Q-FUNK>
dberkholz: mraudsepp
02:20
<dberkholz>
ah, cool.
02:20
see if he can make it to portland OR in a couple weeks
02:20
that's when i'll be working on it w/ vagrantc
02:21
Q-FUNK: btw your xorg-announce mail still shows up without a real name..
02:21
<Q-FUNK>
might be hard as he just barely got hired and is currently working on kernel issues. besides, the US embassy doesn't deliver visas THAT fast.
02:21
dberkholz: blame mailx.
02:22
<dberkholz>
Q-FUNK: i don't care about blame, i'd just like it to be fixed like the rest of 'em. =)
02:22
<Q-FUNK>
dberkholz: you're welcome to send a patch. :)
02:23
<dberkholz>
to what? everyone else's stuff works, it can't be a general problem
02:24
<Q-FUNK>
the release generation script is borken. it signs the whole message, which prevents the Subject and To lines to be caught, to begin with
02:24
and I really have better things to do than track down mailx issues.
02:25
anyhow, the release messages is signed by my GPG key, which oughto to be enough.
02:25
<dberkholz>
it's just annoying to see your email show up when everyone else has a name.
02:26
not really a huge deal, but it bugs me every time you bump
02:26
<Q-FUNK>
if it's not a huge deal, then stop bringing it up. I've already said before that I have better things to do than track down mailx issues, so drop it already.
03:10vagrantc has quit IRC
03:20
<daduke>
hi all. Does anybody know how/where the xorg.conf file is generated during LTSP boot? thanks.
03:22exodos has joined #ltsp
03:33
<tarzeau>
GOOD MORNING!
03:57subir has joined #ltsp
03:57OvaKill has quit IRC
03:57F-GT has joined #ltsp
04:40cyberorg has quit IRC
04:48cdealer has joined #ltsp
04:49
<cdealer>
1
04:50cdealer has quit IRC
04:52
<PerfDave>
tarzeau!
04:54cdealer has joined #ltsp
05:01subir has quit IRC
05:06otavio has quit IRC
05:10mhterres has joined #ltsp
05:12
<mhterres>
morning guys
05:27exodos has quit IRC
05:41
<tarzeau>
how can i have an xorg.conf with two more fontpath lines?
06:06ogra has joined #ltsp
06:31Q-FUNK has quit IRC
06:51cliebow_ has joined #ltsp
06:53Q-FUNK has joined #ltsp
07:01wellspinto has joined #ltsp
07:02
<wellspinto>
Hi! I'm need help in ltsp5 with edubuntu 7.10. The Automatic login not work. Apears only X in black window. Any help???
07:05jvanrooyen has joined #ltsp
07:07jammcq has quit IRC
07:13
<wellspinto>
Hi! I'm need help in ltsp5 with edubuntu 7.10. The Automatic login not work. Apears only X in black window. Any help???
07:15
<cliebow_>
wellspinto, i havnt heard any discussion of it...perhaps ogra will chime in..
07:21
<wellspinto>
tks
07:22wellspinto has quit IRC
07:23
<jvanrooyen>
Hi all I'm having problems with tftpboot...using edubuntu and get PXE:E32 timeout, can anyone help with this?
07:42sakhi has joined #ltsp
08:03cdealer has quit IRC
08:07
<jvanrooyen>
Hi all I'm having problems with tftpboot...using edubuntu and get PXE:E32 timeout, can anyone help with this?
08:15sakhi_ has joined #ltsp
08:20sakhi_ has quit IRC
08:22Gadi has joined #ltsp
08:25mikkel_ has quit IRC
08:26nicoAMG has joined #ltsp
08:29sakhi has quit IRC
08:57mikkel_ has joined #ltsp
09:00
<jvanrooyen>
Hi all I'm having problems with tftpboot...using edubuntu and get PXE:E32 timeout, can anyone help with this?
09:04
<tarzeau>
PerfDave!
09:05Guaraldo has joined #ltsp
09:13DonSilver has joined #ltsp
09:23spectra has joined #ltsp
09:23cje2176 has joined #ltsp
09:24
<cje2176>
hi there. I've got a problem using localdev
09:25
when I try to mount the device by hand using the ssh -X /usr/sbin/ltspfsmounter /tmp add command it says that /media/myuser isn't mounted
09:25
I looked on the server and the directory exists and contains device-named directories
09:26
but if i issue an ls -l here's what i get :
09:26
<Gadi>
is your user a member of the 'fuse' group?
09:27
<cje2176>
Gadi: yes
09:27
?--------- ? ? ? ? ? floppy0
09:27
?--------- ? ? ? ? ? tmp
09:27
?--------- ? ? ? ? ? usbdisk-sda1
09:27
that's as root user
09:27
<Gadi>
right - root cannot see into the mounts
09:27
<cje2176>
as the normal user I get correct owner/perms/etc
09:27
<Gadi>
try the ls -l as that user
09:27
<cje2176>
ok
09:28
<Gadi>
yes - our localdev security is such that no other user - even root- can access the mounts
09:28
<cje2176>
ok
09:28
<jvanrooyen>
Gadi: Hi I'm having problems with tftpboot...using edubuntu and get PXE:E32 timeout, can you help with this?
09:28
<Gadi>
jvanrooyen: what type of switch do you have?
09:29
<jvanrooyen>
Gadi: Planet
09:30
<Gadi>
jvanrooyen: is it managed or unmanaged?
09:30
<jvanrooyen>
Gadi: unmanaged
09:30
<cje2176>
Gadi: hum in fqact it's working :)
09:31
<sutula>
Gadi: Can you point me to the mechanism that gets local devices to show up on the gnome desktop? They're in /media/<username>/<device>, correct owner, but not on the desktop.
09:31
<Gadi>
jvanrooyen: http://www.bootix.com/support/problems_solutions/pxe_e32_tftp_open_timeout.html
09:31
<cje2176>
Gadi: do you know if it's possible to have the same behaviour in kde ad when you plug in a usb key on a "normal" workstation (a dialog asking what to do)
09:32
<jvanrooyen>
Gadi: I'll have a look at that now thanks
09:32
<Gadi>
ok, ok one at a time ;)
09:32
<cje2176>
:)
09:33
<Gadi>
sutula: gnome-vfs, I believe, in Ubuntu (I dont think we create .desktop icons anymore, we let gnome do the icon creation - but I have to check)
09:35
cje2176: you can modify /usr/sbin/ltspfsmounter on the server (its in python) - or find out how KDEs mechanisms work - I am not too familiar
09:35
<cje2176>
ok
09:35
thanks, i'll look around
09:35
<Gadi>
but, basically udev on the client will call ltspfsmounter on the server
09:35
so, you can have it pop up whatever you like?
09:35
er like :)
09:35
<laga>
what does ltspfs do? it doesn't happen to provide space where a client can write persistent changes? :)
09:35
<cje2176>
ok
09:40
<Gadi>
laga: ltspfs is our network-based automounter
09:40
<laga>
Gadi: ah, ok.
09:41
<sutula>
Gadi: So ltspfsmounter gets things mounted in /media, then the gnome-vfs-daemon "notices" and puts something on the desktop?
09:42
<Gadi>
sutula: pretty much - afaik
09:42
<sutula>
Thanks!
09:43* sutula suspects there's a vfs-daemon error log somewhere
09:43
<Gadi>
I think it will do that for anything in /media
09:43
prolly
09:43
but, afaik, we don't trigger anything explicitly in the code
09:43
unless it is done in lbmount
09:43
but I doubt it
09:45
bbiab
09:49Nubae has joined #ltsp
09:52
<cje2176>
Gadi: hoz do you handle clean removale of usb keys ?
09:53
s/hoz/how
09:54jvanrooyen has quit IRC
09:57
<sutula>
cje2176: After a few seconds of inactivity, they are safe to remove
09:57
<cje2176>
sutula: it really depends of the amount of data you copy to it and the speed of your usb bus
09:58* sutula repeats "inactivity"
09:58
<sutula>
Detecting that depends on the device :)
09:58Q-FUNK has quit IRC
09:59
<cje2176>
yeah well I would rather have a button to eject the device safely
09:59
instead of explaining to users : please wait until your drive doesn't blink anymore (assuming it has a led and it blinks when there's activity)
09:59* sutula is just repeating what's been said here before...is not involved in the code to do it...submit a patch :)
10:00
<cje2176>
:)
10:01
<sutula>
Basically, the devices are only mounted during activity, and unmounted a few seconds after there's been none. I think the trick is for the user to know when a device is inactive.
10:02
<cje2176>
sutula: you meen ltspfs unmounts them when they're inactive ?
10:02
<Gadi>
exactly - even if there were a button, if it takes time to flush the queue - you will still have users who pull the stick while its blinking
10:02
yes
10:02
the steady-state of ltspfs is unmounted
10:03
<cje2176>
Gadi: ok
10:03
well then I guess it's enough for my needs
10:03
<Gadi>
users will be users - and no number of buttons will change that :)
10:03
<sutula>
It would be nice if there were a gnome (or other) "USB activity light" widget
10:04
<Gadi>
we could prolly tie a widget to ltspfs that indicates when the mount is mounted or unmounted
10:04
:)
10:05
<sutula>
Many have asked for the "unmount" to work...maybe that could be tied to a check-and-wait-if-mounted, to ltspfs
10:06
<cje2176>
On my desktop machine if i umount a disk with data that needs to be writtrn on it gnome tells we that i need to wait and then tells me that i can remove my usb disk
10:06* PerfDave uses the "Safely Remove" feature in KDE
10:07
<Gadi>
right - it would prolly be only a matter of writing a script with zenity and ltspfsmounter <MOUNTPOINT> remove
10:07
<cje2176>
another quick question : ldm takes on awfull lot of time to show up, any leads on where that could come from ?
10:08
<Gadi>
cje2176: on which version of LTSP5?
10:08
<cje2176>
the thing is X starts up quickly (i can see the black cursor cross)
10:08
Gadi: 0.128 or so
10:08
<Gadi>
what distro?
10:08
<cje2176>
0.124
10:08
feisty
10:08
<Gadi>
ah, fesity
10:09
fesity's ldm is written in python and is super slow
10:09
gutsy's was rewritten in C to adress this
10:09
er, address
10:09
<cje2176>
is there a backport or so;ething ?
10:09
<Gadi>
python on a thin client is painful :)
10:10
not that I am aware of, but I don't do the ubuntu packaging :)
10:10
<cje2176>
ok :)
10:10
dapper's ldm was in python too right ?
10:10
<Gadi>
yeah - everything up to gutsy
10:11
<cje2176>
ok, cause it runs faster. I'm pretty sure it's not python related
10:11
it takes like 30seconds to show the login screen
10:11
<Gadi>
python loads a bunch of libraries - which over the network is rough
10:11
especially on low-powered clients
10:12
<sutula>
cje2176: There may be more functionality in the newer one...hence more libraries
10:12* sutula doesn't run ubuntu, so doesn't know for sure
10:12
<cje2176>
hum
10:12
ok
10:12
thanks anyway
10:17
<Gadi>
sutula: were you the one who was gonna work on hacking the autologin stuff in ldm a bit more?
10:17
<cje2176>
I can still use gdm with ltsp 5 right ?
10:18
<sutula>
Gadi: I hope to...unless someone else gets to it first :)
10:18
<Gadi>
cje2176: SCREEN_07=startx
10:18
cje2176: but, you wont get local devices
10:18
<cje2176>
hum.. crap
10:18
<sutula>
Gadi: I may have time this Tues or Thu...I can only work on it when working from home
10:18
<Gadi>
gotcha
10:19
<sutula>
Gadi: I've never submitted patches into ltsp, but if I can get something working, I'll ask here
10:20
<Gadi>
sutula: just start with bzr - that seems to be the way we work round here
10:21* sutula whines about "yet another version control system" to learn ...only know rcs, cvs, and svn so far :(
10:22
<Gadi>
heh, whatever - if I can figure it out anybody can, lemme tell ya
10:22
:)
10:23cje2176 has quit IRC
10:25exodos has joined #ltsp
10:26
<cliebow_>
Gadi:got a sec?
10:26
<Gadi>
for you, cliebow_? anytime
10:26
<cliebow_>
8~)
10:27
next if ($entry->get_value('ou')) = "People"; just want to skip People..
10:29
<ltsppbot>
"cliebow" pasted "sub to fill @classes with group info" (13 lines) at http://pastebot.ltsp.org/426
10:31
<cliebow_>
if too obscure...no worries..
10:31misui has quit IRC
10:32
<Gadi>
push(@classe,
10:32
typo
10:32
<cliebow_>
not really..
10:32
just a chuckoism'
10:32
<Gadi>
ah
10:33
<cliebow_>
just cant figure out how to skpi certain ou
10:33
<Gadi>
scalar ($entry->get_value('ou')
10:33
so, you want a number?
10:33
dont you want a name?
10:34
<cliebow_>
err..a name..butr that piece works
10:34* cliebow_ cliebow looks up scalar
10:34Blinny has joined #ltsp
10:34
<Gadi>
oh!
10:34
next if ($entry->get_value('ou')) = "People";
10:34
thats wrong
10:34
you want:
10:35
<cliebow_>
drm roll!!!
10:35
<Gadi>
next if ($entry->get_value('ou') eq 'People');
10:35
<cliebow_>
Ahhhhhh!!
10:35
Ill give that a whirl...thanks a pile..!
10:35
<Gadi>
np
10:36
<cliebow_>
perfect!!!
10:44staffencasa has joined #ltsp
10:47DonSilver has quit IRC
11:04sepski has joined #ltsp
11:12jammcq has joined #ltsp
11:12
<jammcq>
hey kidz
11:14
<Gadi>
jammcq!!!
11:14
where the heck have u been?
11:14
<jammcq>
right here in beautiful Michigan
11:15
<Gadi>
ah, no wonder I lost u :)
11:15
<jammcq>
and now you've found me :)
11:37lgbr has left #ltsp
11:41DonSilver has joined #ltsp
11:44exodos has quit IRC
11:46lns has joined #ltsp
11:53vagrantc has joined #ltsp
12:00pdjbarber has joined #ltsp
12:11joebaker has left #ltsp
12:13
<lns>
has anyone used Ubuntu/LTSP5 with VMWare server and Win2k3?
12:15
<Blinny>
Where can I read up on the status of LTSP in Hoary?
12:15
er wait
12:15
the next one. Hardy.
12:16
<lns>
lol
12:16
Blinny, have you checked the wiki?
12:16* lns isn't sure but thinks it might be there if anywhere
12:18
<Blinny>
@ltsp.org you mean?
12:19
<lns>
wiki.ubuntu.com if you're looking for ubuntu status
12:21
i dunno i can't find it myself...
12:24
<Blinny>
I'm ready a post by ogra on nabble. No worries, just curious.
12:31
<vagrantc>
my guess is, hoary being a long-term-support release (i think), it will probably focus on bug fixes from what's in gutsy
12:31
er
12:32
hardy
12:32
heh
12:32
<stgraber>
main change is moving LTSP from Edubuntu to Ubuntu (Alternate CD), that'll be the main change in Hardy
12:32
<vagrantc>
ah, right
12:33
<stgraber>
then fixing bugs and the management UI (UI to create/manage compressed chroots)
12:33
<cliebow_>
so it must be time to mirate to hairy..right?
12:33
so it must be time to migrate to hairy..right?
12:34* vagrantc has always thought heron's looked a little frail
12:34
<Blinny>
Will chroot updates be incorporated in Synaptic by Hardy/Hoary/Hairy?
12:35
<cliebow_>
them legs..them legs..them chicken chicken legs..
12:35
<Blinny>
I guess I'm just looking for a roadmap
12:35laga has left #ltsp
12:36
<cliebow_>
Blinny: i am quite sure the technique will remain the same...
12:36
<stgraber>
Blinny: roadmaps are basically the specs we've been working on at Boston
12:38
<Blinny>
Right-O
12:39
<stgraber>
https://blueprints.edge.launchpad.net/sprints/uds-boston-2007
12:39
<Blinny>
I'm shooting for Happy-release-date to switch from Fedora to Ubuntu for my two installations.
12:39
Thanks stgraber
12:40
<cliebow_>
so hardy is table enough to move one machine over??
12:42
<stgraber>
I'm running Hardy on my laptop for quite some time now (the first Alpha) and it's really stable (>1 week uptime, last reboot being a kernel upgrade) but you'll have all the problems that come with a devel release (broken deps mainly)
12:43
I haven't played with LTSP myself (lack of time) but IIRC it's just an option you have to pass to the alternate CD to start a ltsp server install
12:44
<Blinny>
Yeah. That's what I read also.
12:47
<cliebow_>
stgraber: great: has xorg proceeded beyond the point where nx still functions..as i depend on that for remote admin..
12:50vagrantc has quit IRC
12:50vagrantc has joined #ltsp
12:52
<stgraber>
cliebow_: I don't use NX, do you have a quick way of testing that (without installing NX) ?
12:54
<cliebow_>
i should grab it by the horns i guess..and try it..
12:54Gadi has quit IRC
12:55
<stgraber>
cliebow_: VNC works fine, not sure if that helps though
12:55gentgeen__ has quit IRC
12:55
<cliebow_>
heh..not sure either..i just got everything working about the way it ought to..so may as well break it again...8~)
12:56
takes all the fun out of it when things just work
12:56
<stgraber>
:)
12:57mhterres has quit IRC
12:59jbrett has joined #ltsp
13:05gentgeen__ has joined #ltsp
13:09gentgeen__ has quit IRC
13:11
<lns>
stgraber, is LTSP going to be included in Hoary's alternate cd ala Xubuntu style setup?
13:18
<Blinny>
lns: Yah - http://www.nabble.com/ubuntu-LTSP-move-done-...-testers-wanted-td14952192.html
13:22franciosi has joined #ltsp
13:23
<franciosi>
PING 1201548223 916323
13:23franciosi has left #ltsp
13:24pdjbarber has quit IRC
13:26
<lns>
Blinny, wow that's great news!!! =)
13:26
I <3 consolidation =)
13:50bobby_C has joined #ltsp
13:53
<warren>
Somebody have a Edubuntu or Debian box?
13:54
<vagrantc>
warren: got debian
13:54
<warren>
vagrantc, what does "locale" output for you?
13:54
we have a pastebot here?
13:54
<vagrantc>
!pastebot
13:54
<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.
13:55
<warren>
vagrantc, going to use "locale" instead of /etc/environment, it might be a bit more distro neutral
13:55
vagrantc, but I need to be sure it outputs the right thing
13:55
<ltsppbot>
"vagrantc" pasted "locale on debian etch" (15 lines) at http://pastebot.ltsp.org/427
13:56
<vagrantc>
i *think* ubuntu doesn't have the .UTF-8 bit
13:56
ubuntu assumes full UTF-8 support... i think
13:56
<cliebow_>
yeah it does
13:56
looks identical to yours....
13:57
LC_COLLATE is different
13:57
<vagrantc>
well, that's because i set LC_COLLATE manually
13:57
<cliebow_>
LC_COLLATE="en_US.UTF-8"
13:58
<vagrantc>
i'll give it a whirl on a default install
13:59
warren: what are you using "locale" instead of "/etc/environment" for?
14:03
<Blinny>
LC_COLLAGE="en_US.UTF-8" on default ubuntu
14:03
er, COLLATE
14:03
<vagrantc>
ditto for default on debian
14:03
well, at least with my install
14:03
<Blinny>
I just happen to be setting up a x86_64 ubuntu gutsy ltsp test right now
14:07
<warren>
vagrantc, locale doesn't have LANG=?
14:07
this isn't good
14:07
vagrantc, ldminfod reads /etc/environment
14:07
vagrantc, we don't have that
14:08
vagrantc, ldminfod is also broken in that it runs "locale -a" which has 17 things on ubuntu but 696 things on Fedora.
14:08
<vagrantc>
warren: locale does output LANG= ...
14:08
<warren>
vagrantc, ok then
14:08
vagrantc, I'm editing ldminfod to run locale
14:08
<vagrantc>
it's the very first thing it outputs ...
14:08
<warren>
get LANG= from locale instead of /etc/environment
14:09
<vagrantc>
warren: why is it broken that it reports all possible locales?
14:09
<warren>
vagrantc, ldm's greeter blows up when you have a huge list
14:09
<vagrantc>
warren: ok.
14:09
<warren>
vagrantc, and you really don't want to display that many options to the user
14:09
vagrantc, and that list continues to grow quick in fedora
14:09
vagrantc, we need a rethink of which languages to display, gdm upstream is rethinking it as well.
14:10
<vagrantc>
on debian and ubuntu, the sysadmin decides which languages are supported... though in different ways
14:10
<warren>
vagrantc, for defaultlocale do you agree "locale" is OK?
14:11
vagrantc, strace locale
14:11
vagrantc, does it open /usr/lib/locale/locale-archive on your system?
14:12
<vagrantc>
warren: yes
14:13
<warren>
I think this might work
14:13* vagrantc wonders if warren could be more verbose
14:14
<warren>
vagrantc, /etc/environment is not distro neutral
14:14
vagrantc, instead I want to run locale and get LANG= from that
14:14
<vagrantc>
warren: yes, that much is clear
14:14
<warren>
it seems to do the right thing, right?
14:15
vagrantc, a separate problem is locale -a used by another part of ldminfod
14:15
vagrantc, the list is too long
14:15
we have to rethink how to do that
14:15
<vagrantc>
warren: got it. just wanted to make sure which of the two issues you were talking about
14:16
<warren>
vagrantc, ltsp-fedora tree I've been keeping in sync with ltsp-trunk
14:16
vagrantc, maybe we should just merge it?
14:16
vagrantc, ogra didn't see anything bad in ltsp-fedora
14:16
vagrantc, although we still have to make many changes...
14:17
<vagrantc>
"locale -a" on many debian systems will likely only output 3 things: C, POSIX and whatever locales the sysadmin selected.language and/or region you selected
14:17
<Blinny>
For a LTSP5 server based on gutsy is it preferable to do a server .iso install and then 'apt-get install ubuntu-desktop' or do a desktop .iso install and then 'apt-get install ubuntu-server' ?
14:17
(I'm assuming the kernel from the 2nd will overwrite the kernel from the 1st)
14:17
<warren>
Blinny, how does the kernel matter?
14:17
<vagrantc>
Blinny: either should work fine
14:17* warren has the same kernel everywhere
14:18
<vagrantc>
Blinny: you may not even need or want ubuntu-server
14:18
<Blinny>
I suppose I was under the false impression that the -server kernel was different in its workings or scheduler or PAE than -desktop.
14:18
<stgraber>
you just need ltsp-server (-standalone depending on your need)
14:18
<vagrantc>
warren: where's the branch?
14:19
<Blinny>
This machine will also do dhcp, bind, apache2, imap, mysqld etc
14:19
<vagrantc>
Blinny: install whatever you want :P
14:19
<Blinny>
So I should start with -desktop then apt-get install the above service packages individually.
14:19
<vagrantc>
Blinny: "For a LTSP5 server" it shouldn't really matter :P
14:19
<warren>
vagrantc, bzr://bzr.fedorahosted.org/bzr/k12ltsp/ltsp-fedora
14:19
vagrantc, hold a sec, pushing to it now
14:20
<Blinny>
vagrantc: -server is supported 2 more years than -desktop
14:21
<vagrantc>
Blinny: yes, but an LTSP server without desktop applications is like ... a bicycle without wheels
14:23
<warren>
vagrantc, all of our work is being pushed to upstream trees (ldm, ltspfs, etc.) except for ltsp-trunk at the moment
14:23
<Blinny>
The 7.10 metapackage ubuntu-desktop is what is supported until 2009 then, not the original system one installed.
14:25
<warren>
hm... lmdinfod is making some other bad assumption
14:25
session: lines are full of garbage =)
14:27
<vagrantc>
it's no surprise- the vast majority of development was done by folks working with ubuntu and debian- of course we made assumptions.
14:28
and it's valueable to have other folks getting involved, to root out those assumptions and fix them
14:28
which is why i'm excited we're having the hackfest with gentoo :)
14:32
<warren>
will eharrison go to that hackfest?
14:32
<dberkholz>
one thing i wanna make happen there is getting a decent buildsystem finished
14:32
<vagrantc>
warren: i hope so - haven't heard anything though
14:32
<dberkholz>
besides the obvious (gentoo port)
14:32
<warren>
vagrantc, apparently upstream GNOME and KDE use /usr/share/xsessions to define what sessions are available
14:33
vagrantc, and custom sessions made by sysadmins can go into /etc/dmsessions
14:33
vagrantc, they are .desktop files in there
14:34
it will be quite a bit of code to parse it in exactly the same way
14:34
but ldm should parse it in the same way eventually
14:34
hmm
14:35
<vagrantc>
warren: might be able to just spit out the *.desktop files as comma-separated lists ...
14:35
<warren>
vagrantc, it loops over .desktop files, sees if TryExec= can be executed, if so it populates the lists wiht:
14:35
[Name]
14:35
and [Comment] is a tooltip
14:37
vagrantc, what is the output of: /usr/sbin/update-alternatives --list x-session-manager
14:38
vagrantc, /usr/sbin/update-alternatives --list x-window-manager
14:38
<vagrantc>
warren: just a list of paths to the binaries
14:38
<warren>
vagrantc, I need concrete examples
14:38
<vagrantc>
warren: /usr/bin/icewm-session
14:38
<warren>
oh
14:38
ok
14:38
<vagrantc>
but i know ogra has wanted something better, and making use of the *.desktop files is probably a good move
14:39
<warren>
lots of code though
14:39
<vagrantc>
not ridiculous
14:40
<warren>
vagrantc, does Debian have sessions in .desktop files somewhere?
14:40
ls /usr/share/xsessions/
14:40
gnome.desktop kde.desktop
14:40
<vagrantc>
warren: in debian etch, in /usr/share/xsessions
14:40
<warren>
vagrantc, ok good, we at least have the same thing there
14:40
same location
14:41
<vagrantc>
same in sid
14:41
file format seems pretty simple ...
14:42
some .desktop files seem to have Exec with full paths, some without any pathing
14:42
<warren>
vagrantc, TryExec?
14:43* vagrantc wonders what's the difference between Exec and TryExec ...
14:43
<warren>
vagrantc, TryExec needs to be tested for executability
14:43
<vagrantc>
most only have Exec
14:43
<warren>
our desktop team guys said TryExec should be tested
14:43
I suppose if TryExec doesnt' exist, then use Exec instead
14:43
<vagrantc>
right
14:46
<warren>
vagrantc, want to write that code for ldminfod?
14:53
vagrantc, does this mean ldm can pass to ldminfod *any* pathname and binary to run?
14:55Blinny has quit IRC
14:59
<vagrantc>
warren: other way around
14:59
warren: or am i misunderstanding?
14:59
<warren>
vagrantc, hmm
14:59
vagrantc, nm
15:00
<vagrantc>
warren: ldminfod merely provides the list of things that could be executed on the server
15:00
(and list of locales)
15:00
and some sort of load measure ...
15:00
<warren>
for now I hard coded a session in my ldminfod
15:01
vagrantc, can you write parsing of the .desktop files while we work on other parts?
15:01
<vagrantc>
heh. that's one way
15:01* warren trying to figure out a way to handle the locale -a output...
15:01
<warren>
the list is too long
15:01
kaboom
15:01
<vagrantc>
warren: i could easily write the ldminfod side, but i'd have a harder time with modifying ldm itself
15:02mikkel_ has quit IRC
15:02
<warren>
vagrantc, well parsing the .desktop file to get the sessions list doesn't require ldm modification
15:02
<vagrantc>
warren: every single fedora machine supports every single locale that fedora supports?
15:02
<warren>
vagrantc, no
15:02
vagrantc, but locale -a outputs everything glibc supports
15:02
<vagrantc>
warren: why do y'all have so many locales? :)
15:02
ah.
15:03
<warren>
vagrantc, given that glibc's upstream developers work for Red Hat...
15:03
and wrote the locale tool itself...
15:03
<vagrantc>
-a, --all-locales
15:03
Write names of available locales.
15:04
<warren>
In Red Hat it seems to print EVERYTHING glibc knows about
15:04
<vagrantc>
sounds like fedora's locale command is broken, unless that's a debian-specific patch
15:04
<warren>
might be
15:05
<vagrantc>
warren: have you made patches to ldminfod in your ltsp-fedora branch?
15:05
<warren>
vagrantc, only one
15:05
<vagrantc>
ah, got it
15:06
warren: might be good to start working on feature-specific branches ...
15:06
<warren>
vagrantc, ok, the one change I pushed upstream didn't break anyone yet
15:11
<vagrantc>
warren: yeah, but to merge your ldminfod changes without merging everything else ... the cherry picking support in bzr has some weaknesses
15:11
<warren>
vagrantc, ok, we can use a feature specific branch for further changes
15:14
vagrantc, parsing .desktop files instead of using alternatives can go into ldminfod as-is right?
15:14
vagrantc, doesn't require changes to anything else
15:15sepski has quit IRC
15:15
<vagrantc>
warren: it could be implemented, yes.
15:15
<cliebow_>
WOOT! stopped that infernal beeping..
15:15
<warren>
vagrantc, do you plan on working on that part? I need to know what to work on next.
15:15
<vagrantc>
warren: i can probably get that working, sure.
15:16
<warren>
vagrantc, ok, I'll work on figuring out a better way to handle getting a list of languages to the ldm client
15:16
vagrantc, I'll be back in a few hours, meeting
15:16
<cliebow_>
great to see you guys working so well together..
15:16
<vagrantc>
warren: should i merge the change you made to ldminfod ... or just focus on the *.desktop parsing ?
15:17
<warren>
vagrantc, I guess might as well merge that too.
15:17
vagrantc, doesn't hurt anything right?
15:17
vagrantc, lots of other things we should probably merge
15:17
vagrantc, but I need you and ogra to buy in first.
15:18
<vagrantc>
warren: *most* of your changes seem ok to merge upstream, i think.
15:18
<warren>
vagrantc, oh btw, we came to an agreement with our tftp maintainer
15:18
<vagrantc>
warren: on?
15:18
<warren>
vagrantc, our scripts will see whatever tftp-server is using as its base directory and use that.
15:18
vagrantc, that way it can change our scripts don't need to.
15:18
<vagrantc>
warren: good.
15:18
<warren>
vagrantc, Fedora 8 uses /tftpboot while Fedora 9+ uses /var/lib/tftp
15:19
<vagrantc>
warren: which tftp server are you using?
15:19
<warren>
uhh...
15:19
http://www.kernel.org/pub/software/network/tftp/
15:19* vagrantc waits patiently
15:19
<vagrantc>
supports pxe now?
15:19
<warren>
eh?
15:20
<vagrantc>
for years, that didn't support some options required for most pxe implementations
15:20
i've used tftpd-hpa and atftpd instead mostly
15:20
<warren>
vagrantc, seems to work fine for us
15:20
<vagrantc>
ah, good.
15:20
<warren>
vagrantc, we have other tools using PXE regularly
15:20
we need some binary from syslinux to do it
15:21
<vagrantc>
pxelinux.0 or whatever ...
15:21
<warren>
yeah
15:21
<vagrantc>
anyways ...
15:21
<warren>
a solution for the locale -a problem will require changes to both ldminfod and ldm itself
15:21
I'm picking the brain of our two gdm people
15:21
<vagrantc>
good
15:21
<warren>
they're rewriting the language display/selection part right now
15:22
of gdm
15:22
I wonder if ldm client capability can just be a feature of gdm
15:22
we're reinventing the wheel in a few ways
15:22
<vagrantc>
ideally, i'd like to get a human-friendly valuedisplayed to the user in ldm's session selection
15:23
<warren>
vagrantc, yeah, that'll be part of my changes too
15:23
<vagrantc>
warren: we've long hoped that gdm would someday just incorporate ldm functionality
15:23
<warren>
vagrantc, after you get done with .desktop parsing I'll improve on top of that
15:23
vagrantc, your part wont require changes to ldm itself.
15:23
gotta go now
15:23
bbl
15:23cliebow_ has quit IRC
15:27kaos01 has quit IRC
15:42zodman has quit IRC
15:55kaos01 has joined #ltsp
16:04Guaraldo has left #ltsp
16:05Gadi has joined #ltsp
16:23DonSilver has left #ltsp
16:24
<vagrantc>
warren: well, i've got code to parse the desktop files ... not sure if i need to coerce it to feed ldm a full path
16:24
some desktop files provide full paths, some do not
16:27* vagrantc is floundering finding the python equivalent of "which"
16:30nicoAMG has quit IRC
16:41
<warren>
vagrantc, oh
16:41
vagrantc, I just talked with our desktop team again. they said python has a "config parser" or something that can read .desktop files.
16:41
http://docs.python.org/lib/module-ConfigParser.html
16:41mischko has joined #ltsp
16:41
<warren>
no idea how it works yet
16:42
<vagrantc>
i've had to use that before, and it was a nightmare
16:42
<warren>
ok
16:42
good to know
16:42
<vagrantc>
but it may be my own inability
16:44
might be worth using in the future, when we want to start implementing better names and such
16:44
<warren>
yeah
16:44
<vagrantc>
but for right now, we're just pulling out Exec or TryExec
16:44
<warren>
vagrantc, pushed your implementation?
16:45
vagrantc, hmm... I guess python wouldn't successfully run without a full path?
16:46
<vagrantc>
warren: haven't pushed yet - i'm a little uncomfortable sending out a test weather the binary exists or not
16:46
s,sending out,sending without,
16:47
i don't know if ldm will require the full path or not ... could try it, i guess.
16:47
<warren>
ok, I have to commute in 15 minutes
16:48* vagrantc tests it out
16:48
<warren>
vagrantc, "for dpath in os.environ['PATH'].split(os.pathsep): for fname in os.listdir(dpath): ... etc"
16:49
<vagrantc>
warren: yeah, that's what i was looking at implementing ... a good deal more complicated than i would have liked, but whatever
16:50
sometimes the simplest things in shell are painful in python
16:51
seems to work without the full paths just fine
16:51mccann has joined #ltsp
16:52
<warren>
vagrantc, ok good.
16:52
vagrantc, now I only need to deal with the "locale -a" too big issue
16:52
<vagrantc>
but i still would hate to display a session that wasn't really available
16:52
<warren>
vagrantc, but I'll do this outside of the upstream trunk
16:52ogra has quit IRC
16:52
<warren>
vagrantc, well how are .desktop files appearing there if it isn't available?
16:53
<vagrantc>
warren: well, in /etc/dmsessions or whatever ... if the sysadmin creates a custom session
16:54
<warren>
vagrantc, if they create a custom session that doesn't exist, then that's their problem
16:54
<vagrantc>
warren: sure, i guess.
16:55ogra has joined #ltsp
16:55
<warren>
gotta leave in 5 minutes
16:55
ogra, please review ltsp-fedora branch, we really gotta get stuff merged back into ltsp-trunk
16:56
vagrantc, <lmacken> warren: which = lambda x: join(filter(lambda y: x in os.listdir(y), filter(os.path.isdir, os.environ.get('PATH').split(':')))[0], x)
16:57
ogra, bzr://bzr.fedorahosted.org/bzr/k12ltsp/ltsp-fedora
16:57
bbl
16:57
<vagrantc>
subprocess.command('which', FOO)
16:59
hrm... something like that
17:01bobby_C has quit IRC
17:01
<vagrantc>
subprocess.call(['/bin/which',FOO])
17:02Gadi has quit IRC
17:03
<stgraber>
subprocess.Popen(("which",arg),stdout=subprocess.PIPE).stdout().read() ?
17:04
<vagrantc>
i don't actually need the output, just the return code, so .call is good enough
17:10Gadi has joined #ltsp
17:10cliebow has joined #ltsp
17:13mischko_ has joined #ltsp
17:15mischko_ has quit IRC
17:24jammcq has quit IRC
17:37mischko has quit IRC
17:46Gadi has left #ltsp
17:52cliebow has quit IRC
17:52cliebow has joined #ltsp
18:01rjune_ has quit IRC
18:01rjune_ has joined #ltsp
18:29ogra has quit IRC
18:45vagrantc has quit IRC
18:56staffencasa has quit IRC
19:04cliebow has quit IRC
19:11spectra has quit IRC
20:04Egyptian[Home] has quit IRC
20:05Egyptian[Home] has joined #ltsp
20:37mistik1_ has joined #ltsp
20:37mistik1 has quit IRC
20:37mistik1_ is now known as mistik1
22:07zodman has joined #ltsp
22:09kaos01 has quit IRC
22:10kaos01 has joined #ltsp
22:15mccann has quit IRC
22:35Joris has joined #ltsp
22:35jammcq has joined #ltsp
22:38abadger1999 has quit IRC
22:51Joris_ has quit IRC
23:20
<moquist>
jammcq: is sbalneav around these days? I haven't seen him lately.
23:20* moquist hasn't been seen lately, either, to be fair.
23:25
<sutula>
!seen sbalneav
23:25
<ltspbot>
sutula: sbalneav was last seen in #ltsp 1 week, 1 day, 12 hours, 46 minutes, and 36 seconds ago: <sbalneav> I've been hellishly busy
23:25
<moquist>
Heh; yeah, I guess so. :)
23:25
<sutula>
moquist: ^^^
23:25
<moquist>
sutula: thx
23:26* sutula needed the distraction
23:57cyberorg has joined #ltsp