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


Channel log from 19 May 2007   (all times are UTC)

01:53vagrantc has joined #ltsp
02:24jsgotangco has joined #ltsp
02:32vagrantc has quit IRC
02:46jsgotangco has quit IRC
03:08bobby_C has joined #ltsp
03:11bobby_C has quit IRC
03:11bobby_C has joined #ltsp
03:19Sp1 has joined #ltsp
03:32bobby_C has quit IRC
03:32bobby_C has joined #ltsp
03:36sepski has joined #ltsp
03:44vagrantc has joined #ltsp
04:12bobby_C has quit IRC
04:22highvoltage has quit IRC
04:32Sp1 has quit IRC
04:32* vagrantc drops support for sarge
04:42
<vagrantc>
ogra: so, what's in your feisty branch that shouldn't hit mainline?
04:57* vagrantc has an idea ...
04:58
<vagrantc>
ogra: ok, so ... i want to see all that junk we install as part of 000-basic-configuration as package dependencies
04:58
hrm.
04:58
forget it.
04:58
well, i still want that
04:59
but i was forgetting that we need to split up the init scripts into little pieces anyways ... had an idea to handle the init scripts in /usr/lib/ltsp/ltsp_config like we do with LDM ...
05:00
i still want to split ltsp-client into ltsp-client and ltsp-client-core ... though i'm thinking all the individual pieces should go into -core anyways ... so it's easier to enable a specific feature
05:52plamengr has joined #ltsp
05:54plamengr has quit IRC
05:54plamengr has joined #ltsp
07:04
<cliebow>
vagrantc:here you are by yourself....talkling to yourself 8~)
07:05
<vagrantc>
cliebow: thanks for the ... support :P
07:06
<cliebow>
heh
07:07
i still enjoy listening...even if most of the time i havnt reallly a clue..
07:07
<vagrantc>
well, splitting the ltsp-client package into two packages is nearly the last step to make lessdisks feel totally obsolete :)
07:08
well, and splitting ltsp-server into ldm-server and ltsp-server
07:13Egyptian[Home] has joined #ltsp
07:25vagrantc has quit IRC
07:26robbie has joined #ltsp
07:50freet15 has quit IRC
08:11vagrantc has joined #ltsp
08:28
<vagrantc>
ogra: i'm wondering if we should even maintain ltsp.pot in bzr ... we generate it on build, and it always generates conflicts.
08:28
<jammcq>
g'morning vagrantc
08:28
<vagrantc>
jammcq: buenas tardes
08:28
<jammcq>
so... where you at today?
08:29
<vagrantc>
still in madrid
08:29
<jammcq>
enjoying it?
08:29
<vagrantc>
it's a big city, but there's some greenspace to be had, at least :)
08:29
<jammcq>
yeah, I was there back in 2000. beautiful city
08:30
<vagrantc>
one of the best things is the going rate on excellent blue cheeses and olives :)
08:30
<jammcq>
oooh
08:31
<ogra>
vagrantc, i'm fine with a split as long as we keep metapackages around
08:31
<vagrantc>
for 3.30 euros, i got a hunk of cheese that would have easily cost $20-30 US
08:31
ogra: ah, good.
08:32
ogra: i think i can do it in a way that will make everybody happier :)
08:32
<ogra>
we should have ltsp-client-sound and ltsp-server-sound as well
08:32
ltsp-swap-client and ltsp-swap-server
08:32
<vagrantc>
that's getting to be overkill
08:33
i think simply one package with none of the special features enabled, and another will all of them enabled, but still disableable by configuration file.
08:33
<ogra>
i dont mind as long as i can install ltsp-client and it pulls everything in
08:34
that would be the metapackage
08:34
ltsp-client depends ltsp-client-core, ltsp-swap-client, ltsp-client-sound, ltsp-client-printserver etc
08:35
ltsp-client core gives you a barebone system
08:35
*ltsp-client-core
08:35
<vagrantc>
well, how about this ... ltsp-client-core contains all the code, but none of the dependencies ... if ltsp-client-core detects that a feature is available (i.e. appropriate binaries available) it enables it unless configured to do otherwise
08:35
and ltsp-client contains all the dependencies
08:36
<ogra>
and a default config file ...
08:36
<vagrantc>
that way, you can install only ltsp-client-core and whatever relevent features you want, and they're enabled if they're instaleld.
08:36dean_za_ has quit IRC
08:36
<vagrantc>
ogra: i don't see the need for half a dozen metapackages
08:37
<ogra>
well, with the finer grained packaging you are able to maintain it component wise
08:37
<vagrantc>
yeah, i guess.
08:37
<ogra>
we might want to change from pulseaudio to flubbersoundserver or something in ubuntu, while debian wants to keep pulse ...
08:37
this setup would make it a lot eaier
08:37
*easier
08:38
<vagrantc>
ubuntu will surely change to the latest whatever :P
08:38Q-FUNK has joined #ltsp
08:38
<ogra>
right
08:38
<vagrantc>
ogra: so ... i'm looking at the mainline-debian-feisty merge ... it's not so bad other than the translations
08:38
<ogra>
well, it always depends who develops which feature ...
08:39
one is always ahead of the other
08:39
<vagrantc>
ogra: though it's evil difficult to do a feature-by-feature merge
08:39
<ogra>
not with separated branches for the single features (including separate packages)
08:40
<vagrantc>
well, we don't have separated branches for separate features at this point :(
08:40
<ogra>
the overall picture will get more complicated but you have more fine grained control
08:40
<vagrantc>
i'm just trying to get our bzr branches back in sync now ... we'd best be smarter in the future
08:41
<ogra>
yeah, lets try to work off a single centered branch
08:41
<vagrantc>
so... i almost have a branch that has everything merged in with commits messages like "sync with feisty" and "sync with debian"
08:43
ogra: so why have you consistantly diverged from the package priorities we've set in debian? does ubuntu even make use of package priorities?
08:43* vagrantc just wants this merge to be over with to start working on new features
08:43
<ogra>
in d-i i think ... but not beyond that
08:43
feel free to change them
08:44
<vagrantc>
i really don't see that any of the packages should be more than "extra"
08:44
<ogra>
fine with me
08:44
what are they now ?
08:45
<vagrantc>
some are optional, some are extra.
08:45
<ogra>
(sorry i have a slight pressur to finish some classmate stuff atm, cant do much ltsp aside)
08:45
feel free to syncronize that
08:47dean_za_ has joined #ltsp
08:48a5benwillis has joined #ltsp
08:52
<vagrantc>
i honestly don't see where debian uses priorities either, other than simply to report that they're wrong. heh. :)
09:03
<ogra>
well, for the esseintial packages they are important
09:03
<vagrantc>
jammcq: ok ... so ever since we've implemented nbd swap, ubuntu's been using a different port than ltsp 4.x ... which should we use?
09:04
<ogra>
ltsp 4.x indeed
09:04
<vagrantc>
ogra: are essential packages ever something lower than priority important?
09:04
<ogra>
but make sure to change it on both sides ;)
09:04
no, i dont think so
09:04
<vagrantc>
ogra: change the port on both sides?
09:04
<ogra>
yep
09:05
<vagrantc>
ah, server and client. yes.
09:05
<ogra>
indeed
09:05
<vagrantc>
ogra: so we should switch to the port that ltsp 4.x uses?
09:05* vagrantc wondered if it was debian and ubuntu
09:05
<ogra>
we should switch (or add where switching causes headdache) to the ltsp 4.2 syntax completely
09:06
<vagrantc>
ok.
09:06
i'll keep that in mind
09:06
<ogra>
i like the idea of introducing new variables to a completely new implementation, but the doc overhead that causes seems to much
09:07klausade has quit IRC
09:07
<ogra>
at least we didnt manage well to keep up with docs for the new stuff yet
09:07
so lets rater make sure we match the old docs
09:07
<vagrantc>
so, other than translations, i have a fully merged branch.
09:07
<ogra>
yay
09:07
<vagrantc>
mdz suggested there were some translation merging tools ...
09:09
<ogra>
where d we differ ? i'm fine with dropping mine, my big target is anyway to get it into rosetta and let the ubuntu translator crew loose on it
09:09
<vagrantc>
ogra: so just grab the debian translations?
09:09
<ogra>
i'm pretty sure debians translation status is better than ubuntus atm
09:09
<vagrantc>
most or all of them have at least been reviewed by teams and such
09:10
<ogra>
yeah
09:10
<vagrantc>
the spanish one received a lot of commentary :)
09:10
<ogra>
i need to do a proper rosetta export for them
09:10
i didnt see much need beyond the d-i part yet though
09:10
but now ldm will be translateable
09:11
<vagrantc>
translated --help stuff is a good thing
09:11
<ogra>
so there is something in the users face ....
09:11
<vagrantc>
right.
09:11
<ogra>
right, that too
09:11
<vagrantc>
that's the bulk of the translations at this point... --help options for ltsp-build-client
09:12
<ogra>
right
09:13
ltsp-build-client-gtk will bring some new ones in
09:14
<vagrantc>
eeyk.
09:15
<ogra>
heh
09:15
<vagrantc>
did i mention we should ditch maintaining ltsp.pot in bzr ?
09:15
<ogra>
yes
09:15
<vagrantc>
it seems to just attract conflicts and not really provide any real benefit.
09:16
ok, "upstream" gettext dealt with
09:16
now the debconf translations
09:17
so here's the quandry ...
09:17
ubuntu has several more questions that are not translated in debian at all ...
09:18
<ogra>
really ?
09:18
weird
09:18
<vagrantc>
well, i never merged some of the .udeb stuff
09:19
because it's all hard-coded
09:19
<ogra>
ah, right the dhcpd autoconfig stuff i added
09:19plamengr has quit IRC
09:19
<vagrantc>
but since lenny release is likely a long ways off, we can just merge it and fix it later :)
09:20
what would be the ideal thing to do is to merge all the questions that aren't present in debian at all, but keep the debian translations for things that are translated ....
09:20
and then feed it back to the respective review teams
09:21
gah. this is so close it hurts.
09:23
actually, that doesn't look too hard...
09:25
ogra: do you remember miroslav kure ... were they at UDS?
09:25
<ogra>
hmm, no idea ... i'm not good with RL names
09:26muh2000_ has joined #ltsp
09:26
<vagrantc>
i think i met miroslav, and swore i recognized the name ... and here it is in the cs.po file
09:31
ok ... that was way easier than it looked
09:34
well.... looks good to me :)
09:41muh2000 has quit IRC
09:55rstory has joined #ltsp
09:55
<rstory>
is there a bugzilla for ltsp?
09:57
<vagrantc>
rstory: what distro?
09:57
<rstory>
rhel
09:58
<vagrantc>
redhat hasn't switch to ltsp5 yet ...
09:58
<rstory>
but the problem is in ltsp-update-sshkeys..
09:59
<vagrantc>
well, what's the problem
09:59
<rstory>
it has some hardcoded assumptions..
09:59
<vagrantc>
indeed it does
09:59
<rstory>
/usr/local/sbin/ltsp-update-sshkeys: line 6: tempfile: command not found
09:59
/usr/local/sbin/ltsp-update-sshkeys: line 21: $output: ambiguous redirect
09:59
<vagrantc>
so you're trying to install from tarball ?
10:00
rstory: do you have mktemp ?
10:00
<rstory>
yep
10:00
yep
10:00
and rhel uses /var/log/messages, not syslog...
10:06
<ltsppbot>
"vagrantc" pasted "patches for ltsp-update-sshkeys" (53 lines) at http://pastebot.ltsp.org/140
10:07
<vagrantc>
rstory: that ought to fix it ... even if it's a little ugly.
10:07
<rstory>
no worries.. i'd already fixed my local version.. just wanted to make sure it got fixed upstream..
10:08* vagrantc tries to kick upstream, but doesn't know where to kick
10:08muh2000_ has quit IRC
10:10
<vagrantc>
ogra: i'm not proud of the massive merges: http://llama.freegeek.org/~vagrant/bzr/ltsp/vagrant-ltsp-mainline/
10:10
rstory: how did you patch it locally?
10:12
rstory: could also report on ... ltsp-developer@lists.sourceforge.net
10:15
<rstory>
quick n dirty.. s/tmpfile/mktemp/ s/syslog/messages/
10:16
now the real problem.. ubutnu starts to boot, I get a sliver of color in the progress bar, then it drops bask to text mode, reports "can't access tty; job control turned off", and just sits there..
10:17
<vagrantc>
hrm.
10:17
rstory: pxe boot, or etherboot?
10:17
<rstory>
ah.. never mind.. bad path in mount request
10:18
<vagrantc>
don't include the ip address in the root-path entry in dhcpd.conf
10:26maaarek has joined #ltsp
10:30klausade has joined #ltsp
10:32
<rstory>
right.. so now i have the ubutnu login screen.. how do I redirect to my xdmcp server?
10:33
<vagrantc>
in /opt/ltsp/i386/etc/lts.conf set SCREEN_07=startx
10:38
and if your XDMCP server is different than your NFS server, set SERVER= to point to your XDMCP server
10:39gepatino has joined #ltsp
10:43
<gepatino>
hi, anybody had trouble using usb keyboard and mouse on the clients side?
10:44
i'm having a PCI error after loading the kernel, and the only 'strange' thing in the clients are usb keyboards and mouse
10:54gonzaloaf has quit IRC
11:07
<rstory>
ok, I set SYSLOG_HOST, and XDM_SERVER to point to the XDMCP server, and I see syslog packets, but not xdmcp
11:07
i see nbd_client errors in the syslog.. what's that?
11:11hick518 has joined #ltsp
11:12
<hick518>
Hi, I'm Rob. I've never used IRC before, so I apologize in advance if I make a mistake...
11:15
I'm looking for help in creating/editing udev rules and scripts so that I can enable local access to a PCMCIA card reader that's installed in a thin client. Am I in the right place?
11:23gepatino has quit IRC
11:24
<vagrantc>
rstory: nbd is used to do remote swap
11:24
rstory: if your clients have 64+ MB of memory, you could disable it in /opt/ltsp/i386/etc/lts.conf
11:25
hick518: yeah, you're in the right place. haven't don't much with pcmcia and thin-clients though
11:26* vagrantc grumbles at lintian errors
11:27
<hick518>
vagrantc: I have the lspci info and I know which modules need to be loaded (thanks to knoppix), but I don't understand the rules and scripts well enough to make one of my own. By the way, this is a PCI-based reader that accepts a PCMCIA memory card.
11:28
<vagrantc>
hick518: which distro are you using?
11:28
<hick518>
vagrantc: I'm using K12LTSP 5.0 EL. It's based on CentOS 5, and uses LTSP 4.2
11:29
<vagrantc>
no clue
11:31
rstory: so, what comes up on the thin-client ? a grey screen?
11:33
<rstory>
a ubutnu colored login screen..
11:33
<vagrantc>
rstory: could you paste /opt/ltsp/i386/etc/lts.conf to the pastebot ?
11:33
!pastebot
11:33
<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.
11:34
<ltsppbot>
"rstory" pasted "lts.conf" (7 lines) at http://pastebot.ltsp.org/141
11:35
<rstory>
if I log in on vt1 and run "X -query $XDM_SERVER", I get the remote login window..
11:35
<vagrantc>
rstory: add SCREEN_07=startx
11:36
rstory: you're probably going to have a hard time getting LOCALDEV to work.
11:36
<rstory>
i also tried a remote id in the ubutnu login page, as well as local root... both just return to the login screen
11:36
<vagrantc>
rstory: add SCREEN_07=startx
11:36
<rstory>
i don't care about localdev.. that was just in the default file..
11:36
<vagrantc>
ah, yes.
11:37
rstory: well, as long as it doesn't have any SCREEN_NN entries, it's going to run LDM and not try XDMCP
11:37
<rstory>
ok.. but I'm curious as to what the ubutnu login screen on 7 is _supposed_ to do..
11:37
<vagrantc>
it's a display manager that logs in via ssh
11:37
essentially a frontend for "ssh -X"
11:37
<rstory>
aha... ok, i'll go look on the wiki to see if I can redirect that to another host..
11:38
<vagrantc>
but it makes some debian-ish assumtions. to get something working i'd suggest XDMCP
11:38
rstory: add SCREEN_07=startx
11:38
and then it will use XDMCP
11:41gepatino has joined #ltsp
11:49maaarek has quit IRC
11:53J45p3r has joined #ltsp
11:53vagrantc has quit IRC
11:57gepatino has quit IRC
13:00bobby_C has joined #ltsp
13:08ogra has quit IRC
13:24ogra has joined #ltsp
13:30vagrantc has joined #ltsp
13:31hick518 has left #ltsp
14:47
<vagrantc>
http://lists.alioth.debian.org/pipermail/debian-np-devel/2004-November/000098.html
14:47* vagrantc chuckles
14:53maaarek has joined #ltsp
15:44
<vagrantc>
E: ltsp-client-builder udeb: duplicate-fields-in-templates ltsp-client-builder/build-client-opts Template
15:45
gah
15:45
no idea where that came from
15:45
<rstory>
i think i'm on the verge of getting ldm to work to a different server..
15:45
<vagrantc>
oh, you went that route, eh?
15:46
<rstory>
yeah.. i like the idea of having sessions encrypted..
15:59vagrantc has quit IRC
16:54J45p3r has quit IRC
16:54J45p3r has joined #ltsp
17:11gonzaloaf has joined #ltsp
17:14sepski has quit IRC
17:21Q-FUNK has quit IRC
17:25cliebow_ has joined #ltsp
17:27
<cliebow_>
if( !( $day_of_week % 5 ) ){
17:27
$table = $table."<tr>\n" ;
17:27
}
17:27
<jammcq>
$table .= "<tr>\n";
17:27
<cliebow_>
if $day_of_week is less than zero..
17:28
it evaluates and $table is created
17:28
?
17:29bobby_C has quit IRC
17:30
<jammcq>
I can't see why sign would matter to your logic
17:30
if $day_of_week = -5, then $day_of_week % 5 = 0
17:30
!( 0 )
17:31
<ltspbot>
jammcq: Error: "(" is not a valid command.
17:31
<jammcq>
!( 0 )
17:31
<cliebow_>
this calendar thing worls p[erfectly exept when a month starts on SUnday
17:31
<jammcq>
would cause $table to get updated
17:32
<cliebow_>
ok..Thanks..
17:35
it appears that when $day of week is -1 $table is updated
17:35
which makes nop sense
17:38Q-FUNK has joined #ltsp
17:39Q-FUNK has quit IRC
17:39Q-FUNK has joined #ltsp
18:19maaarek has quit IRC
18:25cliebow__ has joined #ltsp
18:26cliebow__ has joined #ltsp
18:53cliebow_ has quit IRC
19:06cliebow__ has quit IRC
20:16str4nd_ has quit IRC
20:16str4nd has joined #ltsp
20:20spectra has quit IRC
20:21spectra has joined #ltsp
21:20ogra has quit IRC
21:20ogra has joined #ltsp
22:22J45p3r has quit IRC
23:07bengoa has joined #ltsp
23:13
<bengoa>
hi guys. Anybody knows about Artificial Inteligency concepts and/or techiniques used on dspam project? It's for a university work