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


Channel log from 12 November 2008   (all times are UTC)

00:00BrunoXLambert has quit IRC
00:00BrunoXLambert has joined #ltsp
00:04* vagrantc notices that jetpipe doesn't ever actually use JETPIPE_ARGS
00:07
<Ryan52>
ooooooooo....tracebacks from bzr.
00:07
<vagrantc>
oh, fun
00:11
<Ryan52>
does bzr have some way to keep all of the "objects" of all of the branches you check out in one place?
00:11
<vagrantc>
yes
00:11
Ryan52: bzr init-repo -h
00:12
Ryan52: bzr init-repo ~/src
00:12
<Ryan52>
okay, that'll work.
00:12
<vagrantc>
Ryan52: and then anything checked out under ~/src will use all the same stuff
00:13
Ryan52: although i tend to keep a sub-dir for each project, since they have nothing in common.
00:15
wonder what it would take to get ~/.dmrc parsing ...
00:15
although i hear rumours that "~/.dmrc is going to go away" ...
00:16
<rjune_>
what is .dmrc?
00:16
<Ryan52>
what does that do?
00:16
hrm. Session=kde
00:16
<vagrantc>
for gdm and kdm, it defines what each user's preferred session is
00:16
would be nice to respect that from ldm.
00:19
<Ryan52>
what does kde mean?
00:19
<vagrantc>
?
00:19
<Ryan52>
what I mean is, when it sees that how does it know what to do?
00:19twb has quit IRC
00:20
<vagrantc>
Ryan52: yeah, that gets to be the tricky part, unless we can rely on Session to actually be a binary from somewheres
00:20
but i don't think we can...
00:21
<Ryan52>
kde is not a program.
00:21
startkde is.
00:22
<vagrantc>
i think it gets it from /usr/share/xsessions/
00:24
<Ryan52>
makes more sense :)
00:24
<vagrantc>
basically, looks for *Exec in /usr/share/xsessions/*.desktop
00:25
another nice thing would be to use the *.desktop name for the session rather than the actualy executable
00:25* vagrantc really hopes /usr/share/xsessions/*.desktop sticks around a while
00:31
<vagrantc>
darn.
00:32
<Ryan52>
what?
00:32
<vagrantc>
i was hoping we could avoid breaking compatibility by tacking the name with another : ... but it's not really treated as colon-delimeted
00:32
delimited
00:32
we'd have to implement something like session-with-name:KDE:/usr/bin/startkde ...
00:33
and also send the old-style session:/usr/bin/startkde
00:33
or just break compatibility
00:34
<Ryan52>
it hasn't been in use long enough to matter. just break it.
00:34
<vagrantc>
it's about to go into a debian stable release.
00:35
<Ryan52>
oh, I didn't know you got that in.
00:35
<vagrantc>
and it's in ubuntu stable release, as well.
00:35
Ryan52: no, i mean the compatibility breaking part
00:37
well, etch thin clients won't be able to log into servers with the new format for "session:"
00:37
when selecting logins ...
00:37
that's actually not so bad.
00:37
er.
00:37
lenny thin clients
00:48try2free has joined #ltsp
00:51vagrantc has quit IRC
00:52BrunoXLambert has quit IRC
00:52BrunoXLambert has joined #ltsp
01:01dirigeant has joined #ltsp
01:17dirigeant has quit IRC
01:50Pascal_1 has joined #ltsp
01:50Pascal_1 has left #ltsp
02:04alkisg has quit IRC
02:44mikkel has joined #ltsp
02:45exodos has joined #ltsp
03:20Q-FUNK has joined #ltsp
03:26ztupxp has joined #ltsp
03:26|Paradox| has quit IRC
03:26ztupxp is now known as |Paradox|
03:55try2free has left #ltsp
04:01dirigeant has joined #ltsp
04:04dirigeant has joined #ltsp
04:39plamengr has joined #ltsp
04:44sepski has joined #ltsp
05:03bassie__ has joined #ltsp
05:03
<bassie__>
hi all
05:04
I got a client configuration problem in ltsp on ubuntu 8.04. clients booted in a wrong resolution so i modified lts.conf to use 1024x768 and rebuild the client and that solved the problem for 95% of tyhe computers
05:05
i only have a couple of special computers with beamers that still boot in 800x600.
05:05
this is the line that i added: X_MODE_0 = 1024x768 65 1024 1048 1184 1344 768 771 777 806 -hsync -vsy$
05:05
X_COLOR_DEPTH = 24
05:10nubae has joined #ltsp
05:21sepski has quit IRC
05:29Q-FUNK has quit IRC
06:23alkisg has joined #ltsp
06:26alkisg has quit IRC
06:26alkisg has joined #ltsp
06:39Blinny has joined #ltsp
06:46dirigeant has quit IRC
06:57mikkel has quit IRC
07:05alkisg has quit IRC
07:19jammcq has quit IRC
07:21Q-FUNK has joined #ltsp
07:26captain_magnus has quit IRC
07:43jstephan has joined #ltsp
07:51cosf has joined #ltsp
07:56chrisinajar|work is now known as chrisinajar
07:56
<chrisinajar>
'mornin all
07:56six2one has joined #ltsp
07:57jammcq has joined #ltsp
07:57
<jammcq>
Goood morning #ltsp
08:00
<stgraber>
hey
08:00
<chrisinajar>
good morning jammcq
08:01
<jammcq>
hey chris and Stefan
08:01alkisg has joined #ltsp
08:04otavio has joined #ltsp
08:07
<Q-FUNK>
hey
08:08
<jammcq>
hey
08:08
Q-FUNK: what city are you in?
08:09
<Q-FUNK>
currently, going back and forth between Tallinn and Helsinki
08:10johnny has left #ltsp
08:12
<chrisinajar>
so i still can't push to my teams trunk and it's wicked annoying :-P
08:12
<jammcq>
Q-FUNK: ah
08:13
sbalneav and I were discussing having a LTSP hackfest somewhere in europe in 2010
08:13
2009 we're going to brazil
08:13
<Q-FUNK>
ok
08:14
well, now that I'm no longer at Artec, my possibilities to travel are somewhat limitted
08:14rjune has joined #ltsp
08:14johnny has joined #ltsp
08:18
<stgraber>
chrisinajar: a quick fix would be: 1) download the branch again in a different directory 2) copy everything from the branch that's all your changes 3) commit 4) push
08:18
<chrisinajar>
stgraber: do you know what caused this to happen though? That's what I'm more concerned with...
08:19
stgraber: it happened once i merged with trunk and tried to commit, everything just broke
08:19
well, my team's trunk that is..
08:20
is there a different command you're supposed to use besides merge to sync up with a trunk that other's are also working on?
08:20
<stgraber>
chrisinajar: probably something getting out of sync, you should always do a "bzr pull" before commiting
08:20
<chrisinajar>
it gets pissed when i do a pull... says i have to use merge...
08:20
I assume that's corralated with said problem :-P
08:22
i got it to pull, still the same problem
08:23
i'll just do the thing you said up there ^^
08:26
I, apparently, just suck at using bzr.. hah...
08:29Gad1 has joined #ltsp
08:31notasysop has joined #ltsp
08:32
<notasysop>
hi all, im back again with a new problem :(
08:32
some clients boot up, go through the loading screen, then halt with the monitor flashing:
08:33
xauth: creating new authority file /var/run/ldx-xauth-ogWcd3411
08:34
<stgraber>
ogra: I'm packaging the current upstream ltsp, then will do some test on it. When it's stable I'll just tag and will have 3 packages to upload to Jaunty.
08:35
hmm, two actually as ltspfs is still the same
08:36Bruno has joined #ltsp
08:36Bruno is now known as Guest75403
08:37
<stgraber>
Guest75403: hmm, is that only your server or are you actually on IRC ? :)
08:39
<Gad1>
notasysop: sounds like you have some configuration that does not allow the Xserver to load
08:40
check your X_* vars in lts.conf
08:40
<notasysop>
i have other clients that load fine, but they are a different model
08:40
<Gad1>
notasysop: right - it will depend on video card + monitor
08:40
<notasysop>
ah
08:41
<Gad1>
stgraber: I saw sbalneav fixed vagrantc's ldm issues last night
08:41
<notasysop>
ok, there are 450mhz with 8mb vid memory, maybe try a more lightweight desktop?
08:41
they are*
08:42
<Gad1>
notasysop: do you force any modes or color dpeths or anything?
08:42
*depths
08:44
<notasysop>
no i have not changed any settings
08:46
i will change some things in lts.conf my boss is calling me away i will bbl thanks for the help Gad1
08:47BrunoXLambert has quit IRC
08:48
<rjune>
!g
08:49
<ltspbot>
rjune: "g" is Gadi!!!!!!!!!!!!!!!!!!!!!!!!
08:54Ahmuck has quit IRC
08:55six2one has quit IRC
08:56
<Gad1>
morning, rjune
08:56
<rjune>
ogra: I thought ubuntu was getting management stuff for ldap
08:56
how goes it?
08:56try2free has joined #ltsp
08:58six2one has joined #ltsp
09:08
<ogra>
rjune, might be, ask the server team, i stay away from ldap as far as i can
09:08
stgraber, we should be able to sync ltspfs
09:08rjune has quit IRC
09:13
<stgraber>
ogra: yes
09:14
ogra: we already have the latest upstream in Ubuntu anyway but Debian did a change for the udev thing that should now work for both
09:14
<ogra>
right
09:15CAN-o-SPAM has joined #ltsp
09:15
<sbalneav>
Morning all
09:15
<stgraber>
moin sbalneav
09:16
<sbalneav>
hey hey
09:16
I managed to solve vagrantc's problems last night with the latest checkin to ldm.
09:16
Was going to ping warren and see if he could try it.
09:17
ogra: BTW, got a feature in to log normally to a file, or if LDM_SYSLOG is set, to the server.
09:17
<ogra>
sbalneav, yeah, saw that, awesome
09:17
<sbalneav>
Tested by both vagrantc and I, works beeeyoutifully
09:17
<ogra>
coool
09:18CAN-o-SPAM has quit IRC
09:18Gad1 has left #ltsp
09:21
<jammcq>
Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
09:21
<sbalneav>
Hey jammcq
09:23jstephan has quit IRC
09:38mccann has joined #ltsp
09:42gentgeen__ has quit IRC
09:43warren has quit IRC
09:47gentgeen__ has joined #ltsp
09:54etyack has joined #ltsp
09:55jcastro has quit IRC
09:56spectra has joined #ltsp
09:57otavio_ has joined #ltsp
09:58otavio has quit IRC
10:02cosf has quit IRC
10:05try2free has left #ltsp
10:07_UsUrPeR_ has quit IRC
10:08_UsUrPeR_ has joined #ltsp
10:10otavio_ has quit IRC
10:16alkisg has quit IRC
10:21
<sbalneav>
Wow, it got real quiet in here.
10:21
Post-lobster hangover blues.
10:21
<etyack>
hey sbalneav
10:21
<sbalneav>
Hey there!
10:21
<etyack>
heard you guys had a great time
10:21
<sbalneav>
We did.
10:22
We would have had a better time if you had been there :)
10:22
<etyack>
you're to kind!
10:23
i would have had a better time to! virus scare, plus huge spam onslaught sucked up my weekend.
10:23staffencasa has joined #ltsp
10:25* sbalneav imagines etyack with a huge broadsword dressed like Conan battling giant cans of spam
10:25
<jammcq>
mmmm, spam and beer
10:25
<etyack>
fried spam and beer +1
10:26
<jammcq>
fried spam with cheese
10:26
<sbalneav>
I got the subterranian homesick post-lobster hangover blues.
10:26
The pump don't work 'cuz the vandals took the handle.
10:26* jammcq went to red lobster last night
10:26
<sbalneav>
heh, pale comparison no doubt.
10:26
<jammcq>
heh
10:27
only because that's where our user group goes after the meetings
10:27K_O-Gnom has joined #ltsp
10:31
<sbalneav>
I can't remember which astronaut said it, but it was something to the effect of "Once you've been to space, you can never go back to the tilt-a-whirl"
10:31
<jammcq>
depression is common among astronauts
10:32
<sbalneav>
So, once you've been to Southwest Harbor, you can never go back to Red Lobster, by extrapolation.
10:32alkisg has joined #ltsp
10:33
<Q-FUNK>
that makes one wonder if sabdfl is a depressed man
10:34otavio has joined #ltsp
10:35
<jammcq>
sbalneav: that would explain my depression :)
10:37
<sbalneav>
I'm still floating on air that I've seemed to fix all Vagrant's problems with my update last night.
10:37
I'm anxiously awaiting warren to see if it works for him too.
10:38alkisg has quit IRC
10:38warren has joined #ltsp
10:39
<warren>
Is file.bos.redhat.com working as a NFS server?
10:39
<stgraber>
sbalneav: I just did some testing packages for Intrepid, trying them now :) (I still need to package ldminfod and the localapp menu though)
10:39
<Blinny>
sbalneav: You mean all the LTSP-related problems....
10:39
*duck*
10:43
<sbalneav>
warren!!
10:43
<warren>
hi, dealing with stuff here, will test it soon
10:44
<sbalneav>
Perfect, thanks man.
10:48staffencasa has quit IRC
10:54
<stgraber>
btw, I've tagged 2.0.17 locally so it's easier for me to package, if it works I'll then push that revision, any problem with that ?
10:54
<warren>
just do it
10:55
<johnny>
stgraber, do you have a ppa for the new ltsp packages? or will it be easy to use them in intrepid?
10:56
warren, do you happen to know when rpmfusion will have fedora10 directories?
10:56
<warren>
johnny: likely devel will become 10
10:56
johnny: http://rpmfusion.org/Configuration rawhide
10:56
<johnny>
aha.. i'll just remove the others..
10:57
not sure why yum can't fail gracefull tho
10:57
and just say it couldn't find everything.. but continue on
10:57
instead of hard failing all over
10:58
<stgraber>
johnny: I'll push them to my PPA
10:58CAN-o-SPAM has joined #ltsp
10:59rjune has joined #ltsp
11:01alkisg has joined #ltsp
11:04gentgeen__ has quit IRC
11:07
<sbalneav>
stgraber: Do you have hardy packages on your ppa?
11:08
<stgraber>
sbalneav: no, only Intrepid and soon Jaunty
11:10gentgeen__ has joined #ltsp
11:10
<warren>
stgraber: did you push the tag? I want to build from that tag.
11:10japerry has quit IRC
11:15staffencasa has joined #ltsp
11:16
<stgraber>
warren: ok, I push it now
11:16
warren: done
11:23
<notasysop>
where might i find lts.conf?
11:26lucascoala has joined #ltsp
11:26
<sbalneav>
notasysop: Are you running Ubuntu?
11:27Morken has joined #ltsp
11:28cliebow has joined #ltsp
11:29
<cliebow>
chrisinajar, hows your project coming along??
11:29
<chrisinajar>
cliebow: lovely. Working on other stuff right now though... Last i checked the reboot button works...
11:29
<cliebow>
awesome!
11:30
<chrisinajar>
yeah, i was pretty pumped
11:31
<notasysop>
yes i am running ubuntu
11:31
8.10
11:31
<chrisinajar>
cliebow: come 5pm i shall continue work on it :)
11:32
cliebow: until then I am at work, so I can't work on non-work things :-P
11:32
<sbalneav>
notasysop: There isn't one by default, but you can easily create one in /var/lib/tftboot/ltsp/i386
11:33
<notasysop>
ok ill take a look there thanks
11:34
<johnny>
is there a way to find out what packages are in intrepid-proposed without adding it ?
11:34
i'd like to see if there are any kernel packages in there..
11:34
i'm pretty upset at this sata read bug..
11:34
and not having sound..
11:35
my fellow collective members are gonna kill me if i don't fix the sound soon
11:35
<notasysop>
i have problems booting a client with 8mb vid cards, is there an easy way to get around this?
11:36
<johnny>
how do you know the video cards are the probelm?
11:36
<notasysop>
they freeze and flash this on the monitor: xauth: creating new authority file /var/run/ldx-xauth-ogWcd3411
11:37
<johnny>
that could be any other number of reasons
11:38Egyptian[Home] has quit IRC
11:38
<johnny>
maybe it isn't that they are 8mb.. but that they are using the wrong drivers
11:38
or maybe the driver has a bug
11:38Egyptian[Home] has joined #ltsp
11:38
<johnny>
i mean.. it could be that they are 8mb.. but that would only be a problem if they were using too high resolution or too great of a color depth by default
11:39
<notasysop>
how can i change that default setting? lts.conf?
11:40
<johnny>
sure..
11:41
but the best thing would be to try to change to the first console and login to it
11:41
and see what is in /var/log/Xorg.0.log
11:41
but you'd need to set SCREEN_02=shell first
11:41
in lts.conf
11:41
that way you know what you're dealing with
11:42
another option would be to set CONFIGURE_X=N in lts.conf and see if X's own config does better detection than ubuntu
11:42
<notasysop>
ok let me take a look at the log real quick
11:42
<johnny>
ultimately we'd prefer if CONFIGURE_X never had to be used..
11:42
but on the kind of hardware many people have, that isn't a possibility quite yet
11:45dtrask has joined #ltsp
11:46
<notasysop>
what should i be looking for in the log file?
11:49
<johnny>
an error :)
11:49
this is on the client's xorg
11:49
not the server
11:51
<notasysop>
how can i get to the clients log file if it freezes at creating authority file?
11:51
<johnny>
you sure you cannot type ctrl + alt + f2 ?
11:51
and get to a console?
11:52
but as i said.. you need SCREEN_02=shell in lts.conf to make it work
11:52
<notasysop>
ok ive never created a lts.conf file
11:52
<johnny>
it jsut needs to have [default]
11:52
on one line
11:52
<notasysop>
should i go find an example file or just make one with that line in it
11:52
<johnny>
and the next line can be SCREEN_02=shell
11:52
<notasysop>
ok
11:52
<johnny>
that's it
11:52
in /var/lib/tftpboot/ltsp/i386/
11:53
<notasysop>
im gonna give that a shot :) thanks jonny i really appreciate the effort in helping me
11:58
ok i have the file created and in the right dir
11:58
i dont have to restart any service for this to start working?
12:00Egyptian[Home] has quit IRC
12:01
<johnny>
only the client
12:01
since it gets the lts.conf on boot
12:01etyack has quit IRC
12:02Egyptian[Home] has joined #ltsp
12:03
<notasysop>
ok now the client just stops at a blinking cursor
12:03dtrask has quit IRC
12:03
<notasysop>
but my client that used to work fine isn't working
12:03
bad file descriptor
12:04
<johnny>
and if you remove the lts.conf ?
12:04
or rather.. move it..
12:04
<notasysop>
trying that now
12:07
ok my client that worked before is working again
12:07
my other is stuck at flashing cursor
12:07
<johnny>
well.. then lts.conf didn't have anything to do with it..
12:07
sounds like you broke lts.conf
12:08
what does it look like
12:08
the one you used
12:08
it should look like simply
12:08
[default]
12:08
SCREEN_02=shell
12:08
<notasysop>
^^^^ like that exactly
12:08
<johnny>
that shouldn't break anything..
12:09Rideh has joined #ltsp
12:09
<johnny>
sure you didn't have any extraneous spaces or something?
12:09* johnny isn't sure how robust the parser is
12:09
<notasysop>
ill try it again
12:10
if i go to system-admin-login window and change setting for remote, will that affect ltsp in any way?
12:10
<johnny>
why would you enable that?
12:11
<notasysop>
i wouldn't normally, just asking
12:11
<johnny>
don't :)
12:11exodos has quit IRC
12:11
<johnny>
unless you know you need it that is..
12:12Blinny has left #ltsp
12:13
<notasysop>
error Ioctl/1.1a failed: Bad file descriptor is the error on the client
12:14BrunoXLambert has joined #ltsp
12:17
<notasysop>
im going to swap video cards and see if that get me anywhere
12:19Morken has left #ltsp
12:20japerry has joined #ltsp
12:22
<johnny>
notasysop, no idea.. maybe wrong permissions ?
12:22
doubtufl..
12:22
but who knows..
12:23
<stgraber>
Are you ok with me tagging a new ltsp as well ?
12:23
warren: ^
12:24
<warren>
just do it
12:25brendan0powers has quit IRC
12:25brendan0powers has joined #ltsp
12:30
<stgraber>
johnny: both are now uploaded to my PPA, it'll take a while to build and will probably not work (and lacks ldminfod and someth other stuff)
12:31
sbalneav: when they'll be stable enough, I'll rebuild for hardy/intrepid and ask for upload to Jaunty
12:31
<warren>
BREAKING NEWS: Troubled Asset Relief Program Will Not Purchase Troubled Assets
12:31
http://lolfed.com/2008/11/12/breaking-news-troubled-asset-relief-program-will-not-purchase-troubled-assets/
12:32chrisinajar is now known as chrisinaLUNCH
12:34
<johnny>
so.. rebuild for inclusion in intrepid?
12:34
i just want to be able to use all the new stuff that happened at the hackfest
12:34
<stgraber>
johnny: it won't be included we can't push new versions to released version of Ubuntu
12:34
<johnny>
yeah.. that's what i thought..
12:34
i'll use it anyways..
12:34
<stgraber>
johnny: but you will be able to use a PPA for that (maybe not mine, but a team PPA)
12:35
<johnny>
yeah.. please make one :)
12:40CAN-o-SPAM has quit IRC
12:44
<notasysop>
ok jonny
12:45
deleted lts.conf, swapped out video cards, now its booting
12:45
rather quick and smoothly w00t
12:49
so trident 8mb agp card = /epicfail
12:49
thanks for the help
12:49gentgeen__ has quit IRC
12:52dmaran has joined #ltsp
12:52gentgeen__ has joined #ltsp
12:53Lns has joined #ltsp
12:56Lns has quit IRC
12:56
<dmaran>
Greetings All - We are looking for a way to rescue dead sessions for clients. What we mean by that is, we want to be able to connect to the session even if it fails at the client side ie. Power failure etc. As we have UPS on the Servers for 10-20 minutes we should, in theory, be able to connect to all left or hung sessions and save pertinent information before the allotted time expires. Thoughts? We have tried iTalc but we lose the connection with
13:00chrisinaLUNCH is now known as chrisinajar
13:01
<stgraber>
sbalneav: around ?
13:01
sbalneav: ldm seems to work but I get a stacktrace when it shuts down before restarting
13:04
<johnny>
dmaran, there are no such facilities that i know of.. there is no "screen" for X
13:05
<stgraber>
sbalneav: http://www.stgraber.org/download/ldm-trace.png
13:06
dmaran: you could use NX but it's not included in LTSP and persistent sessions don't always work. You'll also loose the ability to run localapps and possibly some other things
13:07Lns has joined #ltsp
13:08
<ltsppbot>
"Lns" pasted "Chroot update w/locale errors - is this a problem?" (82 lines) at http://pastebot.ltsp.org/102
13:08
<Lns>
Hey all..any comments welcome on that paste. I need to reboot in a sec
13:10Lns has quit IRC
13:11jc2it has joined #ltsp
13:11
<jc2it>
TTY=$(tty)
13:11
<Gadi> TTY=${TTY#/dev/tty}
13:11
<Gadi> chvt ${TTY}
13:11
Doh!
13:11
sorry wrong info
13:12Lns has joined #ltsp
13:13plamengr has quit IRC
13:17
<jc2it>
Here is a useful link for those of you who like HOWTOS on creative ways to use LTSP. http://ubuntuforums.org/showthread.php?p=6160681
13:23
<Lns>
jc2it: nice
13:24* Lns didn't know you could simply create screen.d/ files and reference from lts.conf, thought they were set values. crafty
13:30
<sbalneav>
dmaran: You'd probably want to use something like vnc
13:30vagrantc has joined #ltsp
13:30
<stgraber>
sbalneav: Any idea why ldm seems to crash on logout ?
13:30
<sbalneav>
stgraber: Hm, ok, I can poke into that. Maybe we've still got a double free at the end somewhere
13:31
What's the last message it logs?
13:31
<stgraber>
sbalneav: I just added ssh to my chroot so I'll be able to give you a bit more debug info
13:31
<vagrantc>
crash on logout?
13:31
<warren>
sbalneav: working here
13:32
<dmaran>
´╗┐sbalneav: would VNC be any different in "hijacking" the session compared to iTalc?
13:34alekibango has quit IRC
13:35
<stgraber>
sbalneav: Shutting down ssh session
13:35
then ldm_spawn: pid = ....
13:35
Process ... exited with status 0
13:35
then it's restarted
13:35
<vagrantc>
so, doesn't the current jetpipe code now assume that all printers are serial printers?
13:38* vagrantc thinks the recent changes in LDM warrant a version bump to 2.1 or some such
13:39
<vagrantc>
as well as probably the changes to ltsp ...
13:39
<warren>
crap, i forgot to make one more change to ltsp-trunk
13:40
<vagrantc>
stgraber: have you already uploaded ldm 2.0.17 and ltsp 5.1.33 ?
13:40wwx has quit IRC
13:40
<sbalneav>
warren: \o/
13:41
<warren>
the change is fedora only
13:41
<sbalneav>
dmaran: Well, X alone can't hijack a session.
13:41
VNC does allow you to connect to an existing session.
13:42
stgraber: OK, let me poke about a bit. It's failing at the end, so no harm done, but I don't like having crashes.
13:42
lets fix it.
13:42Gadi has joined #ltsp
13:43
<sbalneav>
Hey Gadi!
13:43
<Gadi>
!s
13:43
<ltspbot>
Gadi: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
13:43bobby_C has joined #ltsp
13:43
<warren>
stgraber: what process is crashing? when?
13:43
<sbalneav>
ldm, on logout
13:43
After it's shut down the ssh connection.
13:44
<vagrantc>
sbalneav: does it leave a message in the logs?
13:44
<sbalneav>
No, according to stgraber, it's not.
13:45
since ssh_endsession's pretty much the last thing it does, it should be an easy problem to fix.
13:45
<vagrantc>
how do we know it's crashing?
13:45Leolo_3 has joined #ltsp
13:45
<sbalneav>
Ah
13:45
Well, here's one problem.
13:46
<Leolo_3>
HELP! Xlib: Invalid MIT-MAGIC-COOKIE-1 key
13:46
Xlib: connection to "ws254-Cornwall:0.0" refused by server
13:46
How do i fix that?
13:46Q-FUNK has quit IRC
13:46
<vagrantc>
!release
13:46
<ltspbot>
vagrantc: "release" is please mention the linux distro and release you're using :)
13:46
<vagrantc>
Leolo_3: ^^
13:47
<_UsUrPeR_>
ok, serious question here about executing java code on a client with ltsp-localapps
13:47
<vagrantc>
_UsUrPeR_: what's the question?
13:47
<Leolo_3>
server is centos 4.6. ltsp-4.2
13:48
<_UsUrPeR_>
there is some code on a client that needs to be run as follows; usr/lib/jvm/java-1.6.0-openjdk/bin/java server/myserver
13:48
<stgraber>
vagrantc: yeah, tagged both, commited and I published the packages on my PPA, they still need some work though
13:48
<Leolo_3>
server is runnig inside vmware. this vm works "in the field"
13:48
<_UsUrPeR_>
the server/myserver part is specific to the directory that the java command must run from
13:48
<vagrantc>
stgraber: but the upstream tarball is available?
13:48
<Leolo_3>
i copied the VM to a backup server. And now i get that problem
13:48
<_UsUrPeR_>
due to the way the customer set up their java class environment
13:49
how can i use ltsp-localapps to run that on the client
13:49
<Gadi>
Leolo_3: do you get the error on login from the client?
13:49
or do you not even get a login screen?
13:49
<Leolo_3>
not even get a login screen
13:49
<warren>
sbalneav: need a traceback of it?
13:50
<Leolo_3>
xdm error (pid 5847): server open failed for ws254-Cornwall:0, giving up
13:50
<vagrantc>
_UsUrPeR_: ltsp-localapps /usr/lib/jvm/java-1.6.0-openjdk/bin/java server/myserver
13:50
_UsUrPeR_: ??
13:50
<stgraber>
vagrantc: http://ppa.launchpad.net/stgraber/ubuntu/pool/main/l/
13:50
<sbalneav>
Well, there's at least a couple of problems.
13:50
You can't destroy a GHashTable with just a g_free
13:51wwx has joined #ltsp
13:51
<sbalneav>
There's also a host_list that isn't being freed.
13:51
I'll dig into those tonight.
13:51
<Leolo_3>
gabi : shouldn't the ltsp client's X server allow all connections from the server?
13:51
<sbalneav>
We'll have to iterate through the hash table entries and free them individually.
13:51
<vagrantc>
looks like i should wait on new uploads to debian/experimental ... i need to work on other things anyways.
13:52
<_UsUrPeR_>
vagrantc; yeah, it needs to be a specific java release that has some extra .so files in it
13:52
anyway, remove the part with the giant string in front and it would be the same example
13:52
<vagrantc>
_UsUrPeR_: install it on the thin-client, and ltsp-localappsd will happily call whatever commands you tell it to call.
13:52
<Gadi>
Leolo_3: its more like - does xdm allow connections from the client
13:53
<Leolo_3>
Xaccess has * in it
13:53* _UsUrPeR_ will brb
13:53_UsUrPeR_ has quit IRC
13:53
<Leolo_3>
it's listening on 117
13:53
why wouldn't it...
13:53
<Gadi>
Leolo_3: shutdown xdm on the server, find its Xauth file, delete it, and start it back up
13:53
I think the Xauth has a magic cookie for your other vm
13:53
so, you've got a bad cookie
13:54
<Leolo_3>
what does it's Xauth file look like?
13:54
<Gadi>
"stale", if you will
13:54
<sbalneav>
stgraber: yeah, there's a couple of things in there that could cause crashes. I'll fix 'em tonight. I didn't add the ldminfo.c file, so I haven't ever kind of dug into it. I'll audit the code tonight, see what's up.
13:54
<Gadi>
sudo locate Xauth
13:54
I think
13:54
you'll know it when you see it
13:54* Gadi knows little about straight xdm
13:54
<warren>
Does numlockx need to be in the client chroot or server?
13:55
<Gadi>
warren: depends where you run it
13:55
<stgraber>
sbalneav: ok, it's fine. I'll fix the packaging for Ubuntu, I've plenty of time anyway :)
13:55
<Leolo_3>
I only see ~/.Xauthority which obviously isn't the problem
13:55
<Gadi>
it is an Xclient that can be run anywhere
13:55
<sbalneav>
okiedoke.
13:56
I am SO FREAKING HAPPY i've fixed warren and vagrant's troubles, and we've got the dead processes thing taken care of.
13:56
<Gadi>
Leolo_3: do you have a /var/lib/xdm dir?
13:57
gdm puts Xauth in /var/lib/gdm
13:57_UsUrPeR_ has joined #ltsp
13:57
<_UsUrPeR_>
ok, back
13:57
has anyone else had a problem with vmware server stealing the usefulness of the shift and ctrl buttons on your keyboard?
13:58
<vagrantc>
sbalneav: me too!
13:58
<Leolo_3>
gabi : client or server?
13:58
<_UsUrPeR_>
anyway, vagrantc: what I was saying is as follows: is ltsp-localapps the equivalent of entering SCREEN_02 and running the same command?
13:58
<Gadi>
server
13:58
<vagrantc>
_UsUrPeR_: no, it's its own crazy thing
13:58
<sbalneav>
_UsUrPeR_: very, VERY roughly
13:59
Insomuch as in both cases, you're running the process on the thin client.
13:59
<Leolo_3>
gadi : i shutdown xdm. find /var/lib/xdm/authfiles -type f | xargs rm -v ; then restart xdm
13:59
still have the same problem
14:00
/var/lib/xdm/authfiles has 2 files : A:0-wRlhYE Aws254.Cornwall:0-SKYzE0
14:00
<_UsUrPeR_>
Ok, is it possible to view the results of ltsp-localapps once a command has been entered in the GUI? I.E. output of a program in a terminal or something?
14:00
<Leolo_3>
so the MIT-magic cookie is getting from the client -> server
14:00
but when the server uses it to connect back to the client, it refuses
14:00
it
14:01
<Gadi>
Leolo_3: there is one used for the greeter, and a new one created on login
14:01
<vagrantc>
_UsUrPeR_: spawn the localapp in an xterm ...
14:01
<Gadi>
I think
14:01
oh, no
14:01
and 1 for each connection
14:01
<warren>
sbalneav: SIGABRT during logout
14:01
<vagrantc>
_UsUrPeR_: i.e. ltsp-localapps "xterm yourcrazycommands"
14:02
<_UsUrPeR_>
vagrantc: great! :D thx
14:02
<sbalneav>
warren: yeah, sigabrt will be the segfault.
14:02
that g_free's definitely gonna cause a problem.
14:02
<warren>
you found the cuase already?
14:02
<sbalneav>
I'll fix it tonight.
14:03
Yeah, some stuff got added by someone for that whole ldminfo.c file
14:03
<warren>
OK, I can build early tomorrow at the latest
14:03lucascoala has quit IRC
14:03
<sbalneav>
it sets up a list and a hash table, then tries to free them with a simple g_free()
14:03
that won't work.
14:04
I'll dig into it tonight.
14:04
<Leolo_3>
extra data point : i left the client running while i was out for lunch. the login prompt apeared during that time
14:04
<sbalneav>
afk for a bit.
14:04
back on in a sec.
14:04
<Leolo_3>
i tried to log in but it never succeeded
14:04
which smells like a DNS problem
14:04
<Gadi>
sbalneav: I also note that if the ltsp ssh keys are not set up properly, it now errors with "No response from server" rather than "Not authorized..."
14:04
fwiw
14:05
Leolo_3: ah, could very well be
14:05
Leolo_3: is that client ws254?
14:05
<Leolo_3>
yes
14:06
<Gadi>
is the server .254?
14:06
<Leolo_3>
server .2, client is .254
14:06
<Gadi>
and /etc/hosts correct?
14:07
<Leolo_3>
i have /etc/resolv.conf set to use 10.105.0.2, so /etc/hosts shouldn't be an issue
14:07
<Gadi>
well...
14:07
does 10.104.0.2 know ws254?
14:07mattwalstonfromj has joined #ltsp
14:07
<mattwalstonfromj>
Post install for ltsp-standalone failed when restarting dhcpd. I checked log, no mention, attempted to run in foreground as root and get the following message: Can't create PID file /var/run/dhcpd.pid: Permission denied. My system is running on 8.10. Any ideas?
14:08
<Leolo_3>
what do you mean by "know"? Both forward and reverse resolve OK on the server and the client
14:08
<Gadi>
mattwalstonfromj: is eth0 192.168.0.1?
14:08
<Leolo_3>
actually, i don't know how to test IP -> fqdn lookups on the client
14:09
<mattwalstonfromj>
Gadi: no
14:09
<Gadi>
Leolo_3: create a shell on the client and check its hostname and IP
14:09
<mattwalstonfromj>
Gadi: it is 10.0.0.1
14:09
<Gadi>
mattwalstonfromj: then you need to edit /etc/ltsp/dhcpd.conf and then restart the dhcp service
14:09
do not run it by hand
14:09
<Leolo_3>
gadi : i can ping both IPs and both hostnames from the client
14:10
i don't know of the IPs will resolve though
14:10
<mattwalstonfromj>
Gadi: Thanks, I was not aware of the extra config, it was an upgrade and my config was at /etc/dhcpd.conf
14:10
<Leolo_3>
they *do* resolve on the server, which is using the same named so it shouldn't be an issue
14:11
<Gadi>
mattwalstonfromj: dhcpd will honor /etc/ltsp/dhcpd.conf over /etc/dhcpd.conf
14:11
if present
14:11
<mattwalstonfromj>
Gadi: moved to /etc/ltsp/dhcpd.conf.dist and all is fine
14:12
<Leolo_3>
xdm error (pid 7184): server open failed for ws254-Cornwall:0, giving up
14:12
xdm error (pid 6908): Display ws254-Cornwall:0 cannot be opened
14:12
xdm error (pid 6908): Display ws254-Cornwall:0 is being disabled
14:12
<Gadi>
Leolo_3: firewall?
14:13
<Leolo_3>
no
14:13
between the server and client? no
14:13
<Gadi>
do you get a grey screen on the client?
14:13
<Leolo_3>
yep
14:14
<Gadi>
just startx in lts.conf?
14:14
<Leolo_3>
SCREEN_01 = startx
14:14
SCREEN_02 = shell
14:14
<mattwalstonfromj>
Gadi: thanks, all is working
14:14Rideh has quit IRC
14:14
<Gadi>
mattwalstonfromj: np
14:15
Leolo_3: try adding: DISABLE_ACCESS_CONTROL=Y
14:15
and reboot the client
14:16
<Leolo_3>
doing so
14:16
why would that be an issue here, but not "live"?
14:16
<Gadi>
it may be a workaround
14:16
we'll seee
14:16
;)
14:18
<Leolo_3>
nope
14:18
still no go
14:18
<Gadi>
you may want to: nmap -sS ws254-Cornwall
14:19
and see if 6000 is indeed available (which it should be)
14:20
<dmaran>
´╗┐sbalneav: we know X can't hijack it alone, so we went with iTalc and it seemed great. But when the client goes down so does the session according to iTalc, but we still see the connections to the sever-side. What we are now wondering is if VNC uses the same methodology of attaching to a client to get a session or if it can hijack the session at the server side.
14:20
<Gadi>
oh wait -
14:20
Leolo_3: your auth file is ws254.Cornwall , but your hostname is ws254-Cornwall
14:20
sounds like something's screwed up
14:21
dmaran: you need a caching Xproxy like nx
14:22
<dmaran>
Gadi: And this would be exclusively correct?
14:22
<Gadi>
yup
14:22
<dmaran>
Yuck
14:22
<Gadi>
you need something to save state
14:23
only a caching Xproxy will do it
14:24
<dmaran>
Ugh
14:24
<Gadi>
now, when the client dies, its procs should get a SIGKILL
14:24
so, they should die off gracefully-ish
14:24
<Leolo_3>
6000/tcp open X11
14:24
MAC Address: 00:14:2A:E6:A5:C9 (Unknown)
14:24alkisg has quit IRC
14:25ogra has quit IRC
14:27
<dmaran>
Thanks for the information Gadi.
14:27
<Gadi>
dmaran: if your focus is to have apps save data, I would focus on having them die properly
14:28
your session manager will kill them
14:28
try to figure out how
14:28
and make it nice
14:28
if your focus is to reconnect and pick up where you left off
14:28
well, that's when you need state...
14:29ogra has joined #ltsp
14:29
<Leolo_3>
gadi : ummm.... where does xdm get the hostname to use in it's xauth?
14:29
<Gadi>
Leolo_3: dunno
14:29nicoAMG has joined #ltsp
14:29
<Gadi>
gethostname(), I assume
14:30
<dmaran>
Thanks again!
14:30
<Gadi>
maybe it translates - into .?
14:30dmaran has quit IRC
14:30
<Leolo_3>
i assume so
14:30
because ws254.Cornwall doesn't exist anywhere
14:31
<Gadi>
oh, this is ltsp4.2, right?
14:31
make sure that some of the dirs in /opt/ltsp/i386 are symlinks
14:31
hopefully nothing got screwed up there
14:31* Gadi forgets a lot of ltsp 4.2
14:35
<Leolo_3>
on the server DISPLAY=ws254-Cornwall.ltsp.efficom.on.ca:0 xterm
14:35
shows an xterm on the client
14:37alkisg has joined #ltsp
14:38CAN-o-SPAM has joined #ltsp
14:38Gadi has left #ltsp
14:42
<cliebow>
Gadi:if i were thumbing through /var/log/messages and iu suspected hal oe udev was borked what would i look for...
14:44
<Egyptian[Home]>
evening all
14:44
<cliebow>
!
14:44
<Egyptian[Home]>
anybody know a dave fenwick? i got an old email lying around with his name and muekow on it
14:44
<cliebow>
i don't..
14:44
<Egyptian[Home]>
cliebow: :)
14:45
<cliebow>
Hi!
14:45
<Egyptian[Home]>
hello
14:45
<cliebow>
did you see bts photo that wtogami took?
14:45
<Egyptian[Home]>
who me?
14:46
<cliebow>
yeah
14:46
<Egyptian[Home]>
no .. where? why?
14:47
<cliebow>
http://wtogami.livejournal.com/..i am the guy that looks like Jesus
14:47
at the Last Supper
14:47
<Egyptian[Home]>
according to the da vince code .. that wasnt the last supper
14:48nicoAMG has quit IRC
14:48
<cliebow>
yeah..i dont have Mary Magdalene there..either..ragnar wisloff makes a poor substitute
14:48Ahmuck has joined #ltsp
14:49bjs_redhat has quit IRC
14:50
<Egyptian[Home]>
cliebow when you say "look like jesus" who's interpretation are you referring to ? the blonde version or the brown haired version?
14:50* Egyptian[Home] grins evilly
14:50
<cliebow>
didnt know there were two??
14:50
<rjune_>
cliebow: you don't look like jesus
14:51
<cliebow>
ohh yeah??
14:52* Egyptian[Home] has bigger problems than blonde or brown ..how to get my 3year old off the potty!
14:52
<rjune_>
yeah
14:52
you're too clean cut
14:52
<cliebow>
cause i dont have any hair to work with
14:52
<Egyptian[Home]>
cliebow: and the coptics in egypt have him as brown haired .. *nod* with a beard
14:53* cliebow cliebow looks in mirror to see what color his scalp is
14:55
<rjune_>
cliebow: jesus was a hippie
14:56
<Egyptian[Home]>
a hippie before the word was defined ?
14:56
<cliebow>
Gawd this powerpc is maddening..no mouse or kbd ..or ttys
14:56
so was i
14:56
<rjune_>
Egyptian[Home]: he hung out with hookers, lepers, and all manner of "unclean" folks
14:57
many US churches would turn him away
14:57
<cliebow>
the US would turn him away
14:57
for sue
14:57
for sure
14:58* cliebow cliebow's excuse is he is typing over piles of broken ibooks
14:58
<rjune_>
heh
14:59
<Egyptian[Home]>
rjune_ but they were human right?
15:00
<rjune_>
depends on who you ask I think
15:00* Egyptian[Home] is lost ... i need a third opinion
15:01
<cliebow>
going home..laer..!
15:01cliebow has quit IRC
15:01
<rjune_>
bye
15:01
<Egyptian[Home]>
too late
15:01
<rjune_>
Egyptian[Home]: some people like to think they're superior to somebody
15:01
need it almost
15:02
<Egyptian[Home]>
i was reading bash.org this morning .. laughed so hard almost dint make it to work
15:02
rjune .. i believe maslow called it the hierarchy of needs
15:02
<rjune_>
maslow?
15:02
<Egyptian[Home]>
cant remember if superiority was listed per se on that triangle
15:03
http://en.wikipedia.org/wiki/Maslow%27s_hierarchy_of_needshttp://en.wikipedia.org/wiki/Maslow%27s_hierarchy_of_needshttp://en.wikipedia.org/wiki/Maslow%27s_hierarchy_of_needs <- sorry for the wikipedia reference
15:03
just dont quote it anywhere
15:05
<sbalneav>
Back
15:05
<Egyptian[Home]>
and i m off to bed
15:05
gnite all
15:09notasysop has quit IRC
15:12
<warren>
I'm tagging ltsp-trunk in ~1 hour
15:13Ahmuck has quit IRC
15:13
<sbalneav>
Just got an email from that guy who wrote the wy60 screen script
15:14
I replied to him asking him if he'd be willing to let us include it in LTSP. As long as he's willing to license it under GPL2
15:14
or later
15:15
<vagrantc>
what's a wy60 ?
15:16
<sbalneav>
Wyse 60 terminal
15:17
http://home.att.net/~terminals/terminals/ibm-3153.jpg
15:17spectra has quit IRC
15:17
<sbalneav>
ergh, wrong one
15:17
http://www.emmacare.co.uk/Wyse 60.JPG
15:17
I've got one of 'em here at Legal Aid
15:18
<jammcq>
The page cannot be found
15:18
<warren>
why does it need a screen script? what does it do?
15:18
differently
15:18
<jammcq>
damned spaces
15:18
<vagrantc>
it emulates a wy60 ?
15:18
<sbalneav>
starts up a wyse60 emulator rather than Telnet, which is essentially a vt100 emulator
15:19* vagrantc actually messed around with some wyse terminals in the early days of freegeek
15:19
<jammcq>
err, wy60 is nothing like a vt100
15:19
<sbalneav>
right
15:19
<jammcq>
well, sort of like. they were both terminals
15:19
but the similarity ends there
15:19
<sbalneav>
The escape codes for a "Linux" terminal are basically a superset of vt100
15:20
so if you telnet into something, you look sort of like a vt100
15:20
<jammcq>
yes, the 'xterm' and 'console' are nearly vt100 compatible
15:20
close enough for most things
15:20
<sbalneav>
but if you've got an app that was written for a wyse, you're hosed
15:20* jammcq spent a good 15 years knee deep in terminal configurations
15:20
<sbalneav>
that's what his csreen script does.
15:20
<jammcq>
well, hosed, unless you have a wy60 emulator :)
15:21
<sbalneav>
Precicely
15:21
<jammcq>
what's he using for an emulator?
15:21K_O-Gnom is now known as steffy
15:21
<sbalneav>
so, if he's willing to let us include it, someone might find it usefule
15:21
the program wy60
15:21jcastro has joined #ltsp
15:21
<jammcq>
oh, I think that program was written by Markus Gutche
15:21
I know he wrote one
15:22
<sbalneav>
cool
15:22
<jammcq>
same guy that created etherboot
15:22steffy is now known as K_O-Gnom
15:22
<sbalneav>
right
15:22
Speaking of which, Ken Yap pinged you lately?
15:23
<jammcq>
hmm, been a few months
15:23K_O-Gnom has left #ltsp
15:23
<jammcq>
he's on the ltsp-discuss list. he replies once in a while
15:23
<sbalneav>
cool
15:23six2one has quit IRC
15:23warren has quit IRC
15:27vagrantc has quit IRC
15:27daduke_ has quit IRC
15:27Hyperbyte has quit IRC
15:27jbrett has quit IRC
15:27artista_frustrad has quit IRC
15:27Loto has quit IRC
15:27popey has quit IRC
15:27Appiah has quit IRC
15:27Lns has quit IRC
15:27jammcq has quit IRC
15:27alkisg has quit IRC
15:27ogra has quit IRC
15:27wwx has quit IRC
15:27puff has quit IRC
15:27highvoltage has quit IRC
15:27jcastro has quit IRC
15:27Patina has quit IRC
15:27mattwalstonfromj has quit IRC
15:27nubae has quit IRC
15:27|Paradox| has quit IRC
15:27cyberorg has quit IRC
15:27loather-work has quit IRC
15:27F-GT has quit IRC
15:27sbalneav has quit IRC
15:27yanu has quit IRC
15:27ltsppbot has quit IRC
15:27kc8pxy has quit IRC
15:27shogunx has quit IRC
15:27babyhuey has quit IRC
15:27Leolo_3 has quit IRC
15:27rjune has quit IRC
15:27rcy has quit IRC
15:27ccherrett has quit IRC
15:27fgiraldeau has quit IRC
15:27topslakr_ has quit IRC
15:27Lunks has quit IRC
15:27stgraber has quit IRC
15:27CAN-o-SPAM has quit IRC
15:27jc2it has quit IRC
15:27mccann has quit IRC
15:27chrisinajar has quit IRC
15:27mistik1 has quit IRC
15:27oh207_ has quit IRC
15:27monteslu has quit IRC
15:27BrunoXLambert has quit IRC
15:27johnny has quit IRC
15:27bassie__ has quit IRC
15:27sep has quit IRC
15:27vvinet has quit IRC
15:27loather has quit IRC
15:27vlt has quit IRC
15:27xmagixx has quit IRC
15:27primeministerp has quit IRC
15:27sutula has quit IRC
15:27gentgeen__ has quit IRC
15:27brendan0powers has quit IRC
15:27staffencasa has quit IRC
15:27Narcissus has quit IRC
15:27linio has quit IRC
15:27v-pro has quit IRC
15:27japerry has quit IRC
15:27Egyptian[Home] has quit IRC
15:27Guest75403 has quit IRC
15:27lamikr has quit IRC
15:27pscheie has quit IRC
15:27MM2 has quit IRC
15:27topslakr has quit IRC
15:27ninkendo has quit IRC
15:27leio has quit IRC
15:27ltsplogbot has quit IRC
15:27Ryan52 has quit IRC
15:49Ryan52 has joined #ltsp
15:49topslakr has joined #ltsp
15:49sutula has joined #ltsp
15:49primeministerp has joined #ltsp
15:49xmagixx has joined #ltsp
15:49vlt has joined #ltsp
15:49vvinet has joined #ltsp
15:49sep has joined #ltsp
15:49BrunoXLambert has joined #ltsp
15:49Ahmuck has joined #ltsp
15:49bassie__ has joined #ltsp
15:49johnny has joined #ltsp
15:49warren has joined #ltsp
15:49loather has joined #ltsp
15:49pscheie_ has joined #ltsp
15:49brendan0powers has joined #ltsp
15:49chrisinajar_ has joined #ltsp
15:49jcastro has joined #ltsp
15:49CAN-o-SPAM has joined #ltsp
15:49ogra has joined #ltsp
15:49mattwalstonfromj has joined #ltsp
15:49wwx has joined #ltsp
15:49Leolo_3 has joined #ltsp
15:49vagrantc has joined #ltsp
15:49Lns has joined #ltsp
15:49jc2it has joined #ltsp
15:49gentgeen__ has joined #ltsp
15:49japerry has joined #ltsp
15:49Egyptian[Home] has joined #ltsp
15:49staffencasa has joined #ltsp
15:49rjune has joined #ltsp
15:49mccann has joined #ltsp
15:49Guest75403 has joined #ltsp
15:49jammcq has joined #ltsp
15:49nubae has joined #ltsp
15:49|Paradox| has joined #ltsp
15:49rcy has joined #ltsp
15:49lamikr has joined #ltsp
15:49linio has joined #ltsp
15:49popey has joined #ltsp
15:49puff has joined #ltsp
15:49highvoltage has joined #ltsp
15:49MM2 has joined #ltsp
15:49cyberorg has joined #ltsp
15:49loather-work has joined #ltsp
15:49Loto has joined #ltsp
15:49Appiah has joined #ltsp
15:49jbrett has joined #ltsp
15:49daduke_ has joined #ltsp
15:49Hyperbyte has joined #ltsp
15:49artista_frustrad has joined #ltsp
15:49F-GT has joined #ltsp
15:49mistik1 has joined #ltsp
15:49sbalneav has joined #ltsp
15:49ninkendo has joined #ltsp
15:49kc8pxy has joined #ltsp
15:49leio has joined #ltsp
15:49ltsplogbot has joined #ltsp
15:49ccherrett has joined #ltsp
15:49shogunx has joined #ltsp
15:49yanu has joined #ltsp
15:49Narcissus has joined #ltsp
15:49Patina has joined #ltsp
15:49ltsppbot has joined #ltsp
15:49oh207_ has joined #ltsp
15:49monteslu has joined #ltsp
15:49babyhuey has joined #ltsp
15:49topslakr_ has joined #ltsp
15:49stgraber has joined #ltsp
15:49fgiraldeau has joined #ltsp
15:49Lunks has joined #ltsp
15:49v-pro has joined #ltsp
15:50johnny is now known as Guest4935
15:50Ahmuck has quit IRC
15:54Ahmuck has joined #ltsp
15:54Leolo_3 has quit IRC
15:59Guest4935 is now known as johnny
15:59
<sbalneav>
jc2it: Still there?
16:03BrunoXLambert has quit IRC
16:05highvoltage has quit IRC
16:06
<sbalneav>
ltspbot: status
16:06
<ltspbot>
sbalneav: I am connected to freenode as ltspbot.
16:09highvoltage has joined #ltsp
16:11
<sbalneav>
Am I still here
16:12
<warren>
no
16:14
<jc2it>
none of us are here we are were we are but that is not here
16:14
sorry could not resist.
16:14cliebow has joined #ltsp
16:23dirigeant has joined #ltsp
16:24
<sbalneav>
Got your emails.
16:24
We'll get it into the upstream
16:24
I'm heading home, be on later tonight.
16:30chrisinajar_ is now known as chrisinajar
16:31chrisinajar is now known as chrisinajar|work
16:31alkisg has joined #ltsp
16:41jammcq has quit IRC
16:49Guest75403 is now known as BrunoXLambert
16:51warren has quit IRC
16:52CAN-o-SPAM has quit IRC
16:58dirigeant has quit IRC
16:59alkisg has quit IRC
17:03
<stgraber>
Does anyone plan to do some more major changes to the ltsp directory structure or is it safe to work on packaging now ? :)
17:07johnny has left #ltsp
17:17bobby_C has quit IRC
17:17* vagrantc hopes for safety
17:19
<vagrantc>
looks like gadi added the xinitrc.d stuff
17:20
and boy does it seem weird.
17:21
pushing commits while not in irc, too!
17:22lamikr has quit IRC
17:23johnny has joined #ltsp
17:32
<ogra>
xinitrc.d ?
17:35
urgh
17:36
<vagrantc>
yeah
17:36
<ogra>
thats gross, instead of using hal properly for numlock he adds a whole hackish new structure ?
17:36
<vagrantc>
ogra: basically ldm/rc.d/I*
17:36
<ogra>
sigh
17:37
and why doesnt he use rc.d/I* ?
17:38
<vagrantc>
ogra: because he wanted an independent one for rdesktop or ldm or whatever other X related screen scripts
17:39* vagrantc suspects it could've been done very differently
17:39
<ogra>
well, all input handling should properly be done by hal
17:39
now we start to duplicate everything :(((
17:39
<vagrantc>
yes, well, as long as we need to support older stuff, we need some duplication. we all know you'd rather not.
17:40
<ogra>
especially not for new code
17:40
it wont go into gutsy ...
17:41
which is what he wrks with
17:41Q-FUNK has joined #ltsp
17:42
<ogra>
so it really doesnt make sense to pull it into upstream ... with sid debian will support hal input, fedora does it already, so i dont see why that code needs to be added
17:42
instead of working out the proper solution
17:46
<vagrantc>
lenny won't support it
17:46
and i backport current code to lenny
17:47
<ogra>
we have numlock handling in the current code
17:47
<vagrantc>
not in the version in lenny
17:47
<ogra>
its in since ages ?
17:47
iirc it was added with the setxkbmap code
17:48
<vagrantc>
i don't think it got added
17:48
lenny's using 5.1.10 ...
17:49
which i guess was release on 2008-6-25
17:49
<ogra>
yeah
17:49
it was added to ldm back then
17:49
not ltsp
17:49
but well, bloat as you like :/
17:51highvoltage has quit IRC
18:00highvoltage has joined #ltsp
18:15dirigeant has joined #ltsp
18:23Nihilist has joined #LTSP
18:36twinprism has joined #ltsp
18:58
<Ryan52>
okay, time to see if ldm-trunk works for me.
19:01Lns has quit IRC
19:06staffencasa has quit IRC
19:07
<Ryan52>
it works! yay!
19:08* vagrantc cheers
19:12Gadi has joined #ltsp
19:12
<Gadi>
hi, guys
19:12
vagrantc, sbalneav, ogra: ping
19:13
anybody home?
19:13
<vagrantc>
Gadi: heya.
19:13try2free has joined #ltsp
19:13
<vagrantc>
Gadi: wacky xinitrc.d implementation
19:13
<Gadi>
hey - I saw the discussion wrt that
19:13
just to explain a bit
19:13
I only through numlock in there as an example to test
19:14
I know it exists in rc.d/
19:14
<vagrantc>
Gadi: why not just call: xinit ltsp-xinitrc-wrapper ldm $ldm_opts or xinit ltsp-xinitrc-wrapper rdesktop $rdesktop_opts ...
19:14
<Gadi>
six of one
19:14
this way, we can drop a script into the xinitrc.d/ dir
19:15
<vagrantc>
instead, you generate on the fly, a script ...
19:15
yes, i understand that ...
19:15
<rjune_>
Gadi: why you on so late lately?
19:15
<Gadi>
rjune_: I came in to explain myself in front of the jury ;)
19:15
<vagrantc>
but it seems like the dynamic generation is a little wacky ... why not write a script that handles the xinitrc.d stuff, which then calls ldm, rdesktop, whatever
19:16
heh
19:16* Gadi thought that would be protested against more
19:16
<Gadi>
:)
19:16
<rjune_>
naa
19:16
you're proven yourself many times
19:16
<Gadi>
well, in any event - I know that hal will magically handle numlock and other things
19:17
<vagrantc>
Gadi: ogra's on a "why can't everything be done with HAL" crusade ... and it's understandable.
19:17
<Gadi>
this structure will lend itself to more
19:17
or whatever's left
19:17
<vagrantc>
but i think we shouldn't abandon older distros...
19:17
<Gadi>
depending on how you look at it
19:17
<vagrantc>
or something that *can't* be done with HAL
19:17
<Gadi>
well, here's an example:
19:18
<vagrantc>
i guess it can't set the hal properties...
19:18
<Gadi>
say I want to have a smartcard reader and a script that runs in every graphical screen session that requires its presence or it blanks the screen
19:18
now, I can drop said script in xinitrc.d
19:18
and it works for all screen scripts
19:18
I dont see hal doing that
19:19
or say I want to pop up a license agreement before I let you run ldm
19:19
<ogra>
thats definately something you should do with hal and policykit
19:19
<vagrantc>
i'm fully supportive of it, if not some particular implementation details make me unhappy
19:19
<ogra>
all hardware interaction should go through hal on low level
19:19
but i agree there might be toplevel userspace bits you would want
19:19
<Gadi>
what about the xrandr stuff?
19:20
I was thinking that could go here too
19:20
<ogra>
will move to hal on xorg soon
19:20
overrides wont, though
19:20
<Gadi>
overrides?
19:21
as in make my resolution 1024x768 no matter what xorg wants
19:21
<ogra>
but hardware interaction will definately happen in hal/deviceit
19:21
<Gadi>
?
19:21
<ogra>
*devicekit
19:21
no, but having to use hal-set-property to change it
19:22
<Gadi>
hmm... that would be run post-xserver startup, right?
19:22
<ogra>
though that moot with in kernel X anyway
19:22Q-FUNK has quit IRC
19:22
<ogra>
since then you might need a bootparam to even set the resolution
19:22warren has joined #ltsp
19:23
<Gadi>
when does in-kernel X come out?
19:23
<ogra>
i'm not sure how in kernel X handles displays with broken DCC for example
19:24
<warren>
ogra: same way X does, falls back to a sane default
19:24
<ogra>
Gadi, it is in FC10 for certain drivers
19:24
<warren>
and it isn't in kernel X
19:24* vagrantc always likes "good autodetect with option to override"
19:25
<ogra>
warren, well, it happens on boot ... so if you want overrides you will likely needs a boot option or not ?
19:25
<Gadi>
so, it isnt in-kernel? do you mean it is in initramfs?
19:25
<ogra>
i.e. how does it make sure to not fry your VGS tube monitor ?
19:25
*VGA
19:26
Gadi, no, its before nitramfs afaik, warren may correct me here though
19:26
<warren>
no
19:27
kernel mode setting happens when you load a kernel module for your video hardware, that happens in initrd immediately before plymouth draws something after modesetting
19:28
<ogra>
ah
19:28
<Gadi>
and the xserver is run in the kernel?
19:28
<ogra>
parts of it
19:29
<loather-work>
in-kernel X?
19:29
<Gadi>
so, plymouth is an Xclient?
19:29
<loather-work>
this scares me, greatly
19:29
but it could be quite nice at the same time
19:29
<ogra>
loather-work, to late to complain
19:29
already happened
19:29
<Gadi>
hehe - I feel like I just went thru a time warp
19:30
<ogra>
oyu guys really need t read more mailing lists :)
19:30
<Gadi>
guess so
19:30
<warren>
Gadi: what is an example of what you would run in xinitrc.d?
19:30
<Gadi>
well, I guess it is good we pulled X out of ldm
19:31
<loather-work>
haha, if i read mailing lists i wouldn't have time to do my job :(
19:31
<warren>
Gadi: ogra: no, that is not how it works at all
19:31
X does not run in the kernel
19:31
<ogra>
warren, xdandr calls
19:31
*xrandr
19:32
<Gadi>
For now, I would pull most X_ stuff into xinitrc.d: xrandr, dpms, numlock, setkeymap
19:32
<warren>
setxkbmap?
19:32
<Gadi>
until hal implementations exist for some of those things
19:33
sure - why should that be LDM-specific?
19:33
<warren>
Gadi: we need it in both locations with conditionals, because pre-hal distros need to work
19:33
<loather-work>
i dunno, i'm skeptical of this whole move towards HAL
19:33
<ogra>
keymap setting and numlock already works through hal
19:33
<loather-work>
the concept is sane, but the implementation of HAL leaves a bit to be desired
19:33
<ogra>
how about having a ltsp-legacy branch
19:34
<warren>
no
19:34
<stgraber>
ogra: do all distribution currentlz on developement use hal for input devices with X ?
19:34
<ogra>
i'm really tired of bloating the code more and more
19:34
<warren>
We can easily support both pre and post-hal, and with a conditional it'll automatically do one or the other.
19:34
<loather-work>
what i say is have a, "get your distro out of the stone age" philosophy
19:34
<ogra>
we duplicate the world to support obsolete distros
19:34Ahmuck has quit IRC
19:34
<ogra>
warren, with the cost of maintainablity
19:35
<Gadi>
dont get me wrong, btw - I do none of this to support my current gutsy implementation
19:35
<warren>
Your attitude seriously annoys me.
19:35
<Gadi>
I do this work in prep for moving to intrepid or some such
19:35
<ogra>
and yours me, sorry
19:35
but thats a philosophical difference we wont clean up
19:35
<warren>
I need to maintain a single codebase of ltsp that works on RHEL5 and (latest distro) for ~5 years.
19:35
The difference between the two is really tiny
19:35
I will not maintain a separate branch
19:36
<ogra>
for me a released distro doesnt get no new features
19:36
<loather-work>
tell that to the distro maintainers :(
19:36
<warren>
RHEL5 doesn't have LTSP today
19:36
so the LTSP we ship with RHEL5 in the future will be more modernized
19:36
<ogra>
so it should wait until RHEL_&
19:36
*6
19:36
<warren>
You have no right to dictate what I do.
19:36
<ogra>
no
19:36
<Ryan52>
sbalneav: is this patch correct? http://slexy.org/view/s26Klz5HNE
19:37
<ogra>
but thats my opinion
19:37
since it puts a massive load on all of us
19:37
<Gadi>
slexy.org? is that p0rn?
19:37
<warren>
Ryan52: to fix the SIGABRT during logout?
19:37
<Ryan52>
sbalneav: if so, I'll commit and push. (after cleaning up)
19:37
Gadi: no, it's a pastebin :)
19:37
<warren>
ogra: you are severely overstating it
19:37
<Gadi>
riiiiight....
19:37
pastebin in the Ukraine, maybe
19:37
<warren>
Ryan52: i'll test the patch
19:37
<ogra>
you put the maintenance load to support a distro that isnt ready for ltsp5 on all our sholders with that
19:38
<Ryan52>
warren: it's not that. it's the g_free errors thing.
19:38
<warren>
oh
19:38
Ryan52: I thought sbalneav mentioned that the crash during logout is a g_free issue
19:39
<stgraber>
warren: it might be, but I don't think he had a chance to look at it yet
19:39
<warren>
Ryan52: why are we bothering to g_free() everything right before an exit(0)?
19:39
<Ryan52>
warren: I have no clue.
19:39
<warren>
=)
19:40
<ogra>
there is no need to free and glibified vars
19:40
*any
19:40
thast what glib is for
19:40
<stgraber>
so why are we doing it ?
19:41
<ogra>
because the code had free() before sbalneav ported it to glib i guess
19:41
<stgraber>
erm, right
19:41
<Ryan52>
ogra: I know. But you guys seem set in your ways on most things, so I just do by your example and don't argue.
19:41
<warren>
I'm not set in any way
19:41* ogra neither ... apart from curly brackets ;)
19:41
<warren>
I'm wondering why we're going through lots of effort to free stuff right before the process is killed
19:42* Gadi smells a patch that deletes all gfree()'s
19:42
<ogra>
becuae you need to do it that way if you dont use glib
19:42
else you keep fragments in ram
19:42
<Ryan52>
ogra: before the process exists?
19:42
err, exits
19:42
<stgraber>
if these free aren't needed, we should really drop them as they seem to be the source of a lot of problems recently
19:42
<Ryan52>
okay, can I just do that? :)
19:43
sbalneav: can I just do that? :)
19:43
<Gadi>
Ryan52: if it works when you do it...
19:43
<ogra>
right, asl long as you use g_int and g_string etc you wont need to free
19:43
<Gadi>
:)
19:43
<warren>
If you want to really test the effects of g_free() and catch programming errors that otherwise happen only randomly, use
19:43
G_MALLOC_PERTURB=NUMBER
19:43
where NUMBER is a 0-255
19:43
alloc and free then sets a static number
19:44
<vagrantc>
warren: well, i'm glad you're supportive of the efforts to make marginal efforts to keep support for older distros.
19:44
<warren>
vagrantc: I need to maintain an older distro for 5 years, and I don't intend on maintaining a separate branch for only tiny changes.
19:44
<vagrantc>
warren: and i like to maintain support for an older distro too.
19:45
<ogra>
if they are tiny thats fine
19:45
<Gadi>
as long as the lts.conf vars don't change, I'm happy
19:45
:)
19:45
<warren>
only a few conditionals to handle pre-hal and post-hal input setting
19:45
<ogra>
but currently we're discussing to duplicate half the world
19:45
<warren>
what exactly?
19:45
<ogra>
keeping all the pre-hal crap around
19:45
while implementing post-hal stuff
19:45
<warren>
you don't have to ship the pre-hal files
19:45
<ogra>
pulling out developer time
19:46
<warren>
get over it
19:46
I contribute to the latest distro support that you benefit from
19:46
<ogra>
no, but we have to maintain a bloated codebase
19:46
that grows and grows
19:46
<warren>
Other people exist too?
19:46
<Gadi>
can I ask a naive qu?
19:46
<ogra>
i dont deny that
19:47
<vagrantc>
ogra: i think your "half the world" argument is a little exadgerated
19:47
<Gadi>
(well aside from that one)
19:48
would the post-hal stuff be like: if boolean_is_true $SOME_LTS_VAR; then hal_set_property something; fi
19:48
in an init script somewhere?
19:48
<vagrantc>
why not just set up a set of common functions ...
19:48
<Gadi>
I guess it depends, right?
19:49
<stgraber>
not boolean_is_true, it shouldn't require any action from the distros having it
19:49
<vagrantc>
with hal and non-hal variants.
19:49chrisinajar has joined #ltsp
19:49
<warren>
Gadi: no, check for the existence of a post-hal tool
19:49
or something
19:49
<ogra>
Gadi, right, it depends ... and in the end it would be one single looping script that sets the vars
19:49
<Gadi>
but, if it is something that has to happen after Xserver is running, it will still need to be in xinitrc.d?
19:50
<ogra>
vagrantc, because we need to maintain both variants and split developent efforts
19:50
<Ryan52>
warren: oh, looking in the irc logs it looks like they are related
19:50
warren: so ya, test my patch :)
19:50
<vagrantc>
ogra: that's better than requiring folks to fork and loose them entirely.
19:51
if you demand that people comply with something that clearly doesn't work for them, you will loose them.
19:51
<ogra>
at the cost of duplicating a lot of code
19:51
<vagrantc>
some code.
19:51
<Gadi>
wait -
19:51
<warren>
You are seriously overstating it.
19:51
<Gadi>
so, for example
19:51
say I need to change the keymap
19:52
can't my keymap script xinitrc.d just call hal-set-poroperty if such a script exists?
19:52
<warren>
Gadi: hal can happen before or after X starts
19:52
<Gadi>
ah
19:52
<warren>
setxkbmap and numlockx must be after X starts
19:52
<ogra>
Gadi, not in ubuntu
19:52
<Gadi>
thats the piece I was missing
19:52
thx
19:53
<ogra>
it would just read an exported XKBMAP and use the distro script, that happens before hal starts
19:53
<Gadi>
ah
19:53
<ogra>
if yu call setxkbmap that can only work afterwards and is totally moot
19:54
<warren>
Ryan52: could you just push it, with a comment saying "you are not entirely sure this is right"?
19:54
<ogra>
since you can solve the same thing more elegant with a single export
19:54
<warren>
Ryan52: then simple pull for me
19:54
<Gadi>
but, it would still be an LTSP script that would set the override
19:54
just it would do it way before
19:54
<ogra>
well, getltscfg ecports all vars
19:54
*exports
19:55
you dont need any additional script anymore
19:55
the hal handler does it for you
19:55
<stgraber>
ogra: I actually would like to keep the ability of changing the keymap in the xinitrc as in ltsp-cluster the lts.conf can change at any time and I don't like requiring a reboot
19:55
ogra: well, it doesn't have to be in the xinitrc but it has to be somewhere that's called everytime a screen.d script is launched
19:56
<ogra>
you just restart hal before x comes up and export a different value :)
19:56
<Ryan52>
warren: pushed
19:56
<Gadi>
stgraber: just drop a "restart hal" script in xinitrc.d/ ;)
19:57
<ogra>
though i think there is work going on in X to make that work dynamically
19:57
its just not there yet, currently X only reads hal settings at startup if evdev is loaded
20:02
<Gadi>
well, a nice thing about .d directory infrastructures is that a distro can package how ever many of the scripts inside they choose
20:02
right?
20:03
<stgraber>
yeah, it's way better than having the hacks directly in the scripts
20:03
<Gadi>
so, if there is no need for setxkeymap, don't package it
20:03
hows that?
20:03
<warren>
yeah, just delete that file from your package
20:03
<Gadi>
I just eliminated the bloat and the people eat cake!
20:04
<warren>
My rpm .spec would have a conditional to delete it if building for new distros, but keep it for old.
20:04
Maintain a single tree and .spec across all my distros
20:04
<Gadi>
cool
20:04
<warren>
my pidgin.spec supports distros 5 years old until now
20:04
with conditionals
20:06* Gadi exhales
20:06
<Gadi>
:)
20:07
<loather-work>
i switched to carrier recently because of that stupid "feature"
20:07
<ogra>
Gadi, how does that clean the upstream tree
20:08
<loather-work>
not the specfile. the specfile is great :)
20:08
<warren>
what is carrier?
20:09
<ogra>
Gadi, note that i'm not talking for ubuntu at all here, thats stgraber's thing to worry about in the future ;)
20:09
<loather-work>
pidgin with the "you can't resize the text box" feature reverted
20:09
<Gadi>
soeaking of which: stgraber, did you notice the "-s" flag to openvt?
20:10* Gadi wonders if that could be used to specify a default vt when multiple screens are specified
20:10
<stgraber>
Gadi: hmm, X still does it's own switchvt AFAIK
20:11
<Gadi>
yeah, I was thinking that, too, but maybe there's a flag for that, even
20:11
it would be nice - I hate race conditions
20:12
at least it could be used to eliminate the hackish "chvt 7" that exists in ltsp-client-core
20:12
<stgraber>
yeah
20:12
<Gadi>
(talk about yuck)
20:12
warren: dunno if you are aware of that one
20:13
I noticed you moved to VT1
20:13
<ogra>
Gadi, thats a usplash thing
20:13
(the chvt 7)
20:13
<warren>
Gadi: I don't use ltsp-client-core
20:14
<ogra>
it shouldnt be upstream at all, especialy since ubuntu used to use the initscripts policy compliant from the debian dir up to now
20:14
<Gadi>
ah
20:16
<Ryan52>
vagrantc: so I'm about to do the guest login stuff, what should happen if LDM_GUEST_SERVER is not set? then should it just let the user choose whatever (s)he wants?
20:16
<vagrantc>
well, i've continued to maintain them upstream, as they're a potentially useful example, and at least one other distro was using them.
20:17
Ryan52: i think so
20:17
<ogra>
vagrantc, right, but the downstream stuff doesnt really need to be in there
20:17
<vagrantc>
ogra: ??
20:17
<ogra>
upstart should be whitelisted in the RC_ vars in ubuntu 000-basic-configuration, so it uses its own scripts
20:18
<vagrantc>
ah.
20:18
<ogra>
and that code should go from ltsp-client-core
20:18
<vagrantc>
well, upstart still hasn't moved into debian unstable, so i'll still need to use traditional init scripts
20:19
<warren>
my ltsp-build-client supports both traditional and upstart
20:19
since I need to
20:19
<ogra>
vagrantc, sorry, to tired, i meant usplash
20:19
<vagrantc>
ah
20:20jammcq has joined #ltsp
20:20
<jammcq>
hey all
20:20
<Gadi>
hey, jammcq
20:20
bbiab
20:22hanthana has joined #ltsp
20:23chrisinajar has quit IRC
20:28
<Ryan52>
LDM_SERVER should always be set, right?
20:28
<vagrantc>
yes.
20:28
well, from ldm's perspective
20:28
i think the screen script defaults it to SERVER if it's not already set
20:28
<Ryan52>
and it should always be one server, not a list, right?
20:29
or is it a list too?
20:29
<vagrantc>
i thought a space-separated list
20:29
<Ryan52>
ok.
20:29
<vagrantc>
that was added with the load-balancing stuff
20:37
<warren>
Ryan52: btw, using my theme for ldm, there is a weird strip of light blue on the bottom of the screen
20:37
Ryan52: some kind of bug...
20:37
Ryan52: you can reproduce it by using a background that is darker than the standard blue
20:38
Ryan52: your patch fixed the ldm crash
20:40
<Ryan52>
okay, I'll look at it after I finish this, is the theme problem specific to the fedora branch?
20:40
if you don't know, don't bother testing, I will.
20:41
<warren>
Ryan52: I don't know
20:42
Ryan52: http://wtogami.fedorapeople.org/temp/ldm-k12linux-theme-1.0.1.tar.bz2
20:45
<Ryan52>
is there a way to make it reparse lts.conf? as fun as it is waiting for this machine to reboot, no. :)
20:46
<warren>
1962 edition of "The Joy of Cooking" includes instructions for skinning a squirrel.
20:53
<Ryan52>
vagrantc: so, I'm about to push the guest stuff...do you have a testing environment to know if it works? cause I don't. (I know that it doesn't break normal guest login, though).
20:54
<vagrantc>
Ryan52: i've got a test environment, presuming nothing broke since 2.0.17
20:54
nothing else, that is
20:54
though i'm currently working on other projects
20:54
<warren>
I plan on tagging and building tomorrow moring
20:54
morning
20:54
<sbalneav>
Evening all
20:54* vagrantc nudges for ldm 2.1
20:54
<Ryan52>
sbalneav: hi :)
20:55
<sbalneav>
crash bug was the g_free of the hash table?
20:55
I edited it on my laptop, was just going to try it now. Sounds like you may have fixed it?
20:55
<Ryan52>
sbalneav: yup.
20:56
<sbalneav>
Is it pushed up?
20:56
<Ryan52>
vagrantc: okay, well, I ran through the login in my head once, got something to eat, came back and it still made sense...so I'm pretty confident it will work :)
20:56
sbalneav: yup
20:56
<sbalneav>
k, there's a couple of other issues I've fixed as well.
20:56
let me check out what you've done first.
20:57* vagrantc likes good = FALSE;
20:58
<Ryan52>
sbalneav: it's a little messy (should move the variable definitions to the top and rename them)...I meant to clean it up before pushing, but warren wanted me to push so he could test. also, I dunno if there's an easier way for the array than what I did.
20:58
<sbalneav>
yeah, there is.
20:58mccann has quit IRC
20:59
<sbalneav>
gimme a sec, I'll patch it up.
21:03
Any objection to pushing 959?
21:04
<warren>
just do it
21:04
<sbalneav>
k, uno momento
21:05
Done.
21:05
Ryan52: You can use the g_list_foreach, and just recast g_free as a GFunc. It's way coolio
21:06* vagrantc is tempted to re-write Gadi's xinitrc.d stuff a bit
21:06
<Ryan52>
oh, I see.
21:06
<warren>
sbalneav: ogra indicated that you don't need to free things because we are using glib, and I was wondering why go through so much effort to free stuff immediately before exit(0)
21:06* Ryan52 didn't even think there would be a foreach...glib spoils us so much :)
21:07
<sbalneav>
If I allocate it, I free it.
21:07
Good programming style.
21:07
<Ryan52>
that would be what I meant by "set in their ways"
21:07
:)
21:08
<vagrantc>
Gadi: would you be offended if i re-wrote the screen scripts to not dynamically generate an $XINITRC but instead just use xinitrc as a wrapper to ldm/rdesktop or whatever else?
21:09
<sbalneav>
Ryan52: Please point me to where, in any glib documentation, it says you can just end the program without freeing what you've allocated, and I'll gladly remove it.
21:09
<vagrantc>
Gadi: i think it would be considerably less code, and a lot easier to read
21:09
<Ryan52>
sbalneav: hey, ogra said you don't need to free stuff, not me.
21:09
sbalneav: but the kernel isn't going to leave memory allocated to a process that's exited, is it?
21:10
that would be really stupid if it did.
21:10
<ogra>
Ryan52, depends how you allocated it in your proggy
21:10
sbalneav, afaik all g_ vars are handled by glib so you dont need to explicitly free them
21:11
<loather-work>
Ryan52: no, there's no way for a userland process to keep memory claimed once it has exited.
21:11
<ogra>
if you memcpy things around or use lain int instead of g_int etc thats indeed different
21:11
*plain
21:11
<loather-work>
Ryan52: there are exceptions to this, especially with regards to shared memory, but that's an edge case
21:11
<Gadi>
vagrantc: I never take offense at people rewriting my code -> I'm a circuit designer not a programmer ;)
21:12
<vagrantc>
Gadi: heh.
21:12
Gadi: i'll give it a go, then.
21:12
<Gadi>
cool
21:12
<sbalneav>
All I know is, from years of C programming: if you allocate it, free it.
21:12* Gadi will be afk 2moro all day -> goin to meet with the feds
21:12
<vagrantc>
might be something fundamentally flawed in how i'd do it...
21:12
<Gadi>
Ill catch you friday
21:12
<sbalneav>
If I'm running something like Python or Perl, I don't worry. With C, I always do.
21:13dirigeant has quit IRC
21:14
<Gadi>
good night, all
21:14Gadi has left #ltsp
21:16
<ogra>
sbalneav, yeah, i agree thats more elegant and cleaner ... but i really think glib does the magic for you here
21:16
<sbalneav>
Well, the OS will automatically return any alloced memory to the box when ldm exits anyway, which is true. In this case.
21:17
<ogra>
right
21:17
i would heavily disagree if s/glib/xlib ... very bad things happen if you dont free Xlib stuff :)
21:17
<sbalneav>
The PROBLEM is, if you get into the habit of NOT freeing everything you alloc in C, one day, when you write a daemon that runs for a long time, you come into work and wonder why mysimpled is consuming 20 gig of swap :)
21:18
<ogra>
yeah
21:19
<sbalneav>
So, the rule of thumb is: in C, always deallocate, and if you REALLY REALLY have a compelling reason to drop the free's you can think about it carefully, and do so. In our case, it's only gonna save about 15 lines of code, so I'd rather just leave 'em in there. But if someone really wants 'em gone, sure ditch 'em.
21:20
<ogra>
http://tech.shantanugoel.com/2008/05/03/smart-programming-in-c-using-glib.html
21:20
<sbalneav>
yeah, not freeing things in xlib is bad mojo
21:20
<ogra>
Though you still need to be careful about things like freeing allocated memories properly, etc but GLib does ensure that all its functionality is secure. Moreover, GLib has a policy of ensuring that all its functions are threadsafe. This saves you from a lot of checks and balances and locks and scheduling considerations if you had written all this yourself.
21:21
so free is still needed, but if you exit() anyway it will be safe
21:21
<sbalneav>
right.
21:21
<loather-work>
it's g_free() for glib functions no?
21:22
<sbalneav>
So Ryan52's technically right: we don't strictly need 'em. However, we DO have some things on this project that run as daemons (cdpinger, ltspfd*, etc), so, as a personal style point, I think spending an extra 15 lines to "properly clean up after ourselves" kinda keeps us in that frame of mind.
21:23
As opposed to the question that some new future contributor to LTSP may ask: "well, why do you free here and not here?"
21:23
<ogra>
well, Ryan52 just followed my false assumption that glib cares for freeing
21:24
its all my fault !
21:24
<sbalneav>
nah.
21:24
<ogra>
it is :)
21:24
<sbalneav>
Ok, it's all your fault :)
21:24
<ogra>
but its past 4am ... i'm allowed to make mistakes :)
21:24
<sbalneav>
You get a free pass on this one.
21:24
<ogra>
heh
21:25
<sbalneav>
Why up so late and/or early?
21:25
<ogra>
a g_free() pass ?
21:25
<Ryan52>
warren: have you looked at your bg.png lately? running "xli ldm-k12linux-theme-1.0.1/bg.png", I see a light blue bar at the bottom :)
21:25
<sbalneav>
ba-dum-bum
21:25
that's a programming joke, folks
21:25
<ogra>
ne HW in my house ... couldnt leave my fingers off it until it runs
21:25
*new
21:25
but bedtime now
21:26
<sbalneav>
Lullabye, and goodnight, go to sleep little ogra
21:26
<warren>
Ryan52: hahhahaha
21:26
<ogra>
:)
21:26
<warren>
Ryan52: weird, it doesn't show up in gqview, but it shows up elsewhere
21:27
<Ryan52>
it shows up in gqview for me...
21:28
warren: http://ryan52.info/~ryan52/2008-11-12-192652_1280x800_scrot.png
21:39
<loather-work>
anything with 'scrot' in it i'm suspect of :)
21:39* loather-work clicks anyways
21:40
<Ryan52>
'scrot is a commandline screen capture util like "import", but using imlib2.'
21:40
<loather-work>
interesting.
21:40* Ryan52 uses a lot of programs with weird names
21:40
<loather-work>
scrot is also short for a certain part of the male anatomy :D
21:41japerry has quit IRC
21:45
<sbalneav>
loather-work: Well, you need a trailing E for that :)
21:45
wait
21:46
apparently, it can be spelt both ways.
21:46* sbalneav shrugs.
21:46* loather-work nods
21:46
<sbalneav>
I never could bring myself to use BitchX either :)
21:47
<loather-work>
yeah, i'm like that with stupidly-named software as well
21:48
clever names are OK, but i never could answer someone with a straight face, "what irc client is that?" "oh, it's bitch x"
21:49
<Ryan52>
anybody know what ttyname is/was?
21:49
or is it a qt branch specific thing?
21:51
<stgraber>
it's to get the current tty path, it used to be to determine the path to the xauth file
21:52
but it's now using the PID instead in LDM IIRC (maybe it'
21:52
it's used for something else in the qt greeter)
21:52
sorry, not used to the us keyboard :)
21:52
<sbalneav>
I'd like to get my vote in early for the Ubuntu version starting with P
21:52
"Preturbed Pangolin"
21:53
<Ryan52>
what happend to usage? it used to be in ldm.c, and is now gone...
21:54
it's still in ldm.h, though...
21:54
not used anywhere.
21:54
<sbalneav>
Ryan52: Since we don't need any commandline args, we don't need a useage.
21:54
if it's still in .h, ditch it.
21:54
I probably forgot.
21:54
<Ryan52>
what? since when does it not need commandline args?
21:55
<sbalneav>
since the hackfest :)
21:55
<stgraber>
since we don't have X started by ldm anymore
21:55
<Ryan52>
makes sense.
21:55
<sbalneav>
right, ldm runs under X instead of the other way around, so we don't need either the display or tty command line variables anymore.
21:56
Ryan52: You want to fix that, or shall I?
21:56
<Ryan52>
go for it.
21:56
<sbalneav>
http://en.wikipedia.org/wiki/Pangolin
21:57
I'm gonna move around a misplaced comment block while I'm at it.
21:57japerry_cat has joined #ltsp
22:00
<sbalneav>
Pushed 960
22:00
http://en.wikipedia.org/wiki/Greater_Cane_Rat
22:00
Imagine if THOSE infested your house.
22:01
Hmm "Rambunctious Cane-Rat"
22:02
http://www.herorat.org/
22:03CaScAdE^1arAway has joined #ltsp
22:03
<sbalneav>
Explosive sniffing rats. Awesome!
22:03
<loather-work>
they had a special on the discovery channel about those guys
22:03
<sbalneav>
Awesome^2!
22:03
<loather-work>
said they send these rats out in the fields and they start digging on the mines to show where they are
22:04
then people can go and flag them to have them disposed of
22:04
it also showed some of the kids who weren't so lucky, having played soccer or something in a minefield
22:04
land mines are pretty much one of the most evil things ever
22:04
<sbalneav>
Not awesome. Landmines are a big problem.
22:05
Especially since AP mines are specifically DESIGNED not to kill, but horribly maim.
22:05
<loather-work>
yup, 'cause it takes 3 men to care for every wounded one
22:05
or something like that
22:05
<sbalneav>
Sure.
22:06
<loather-work>
either way, war is pretty stupid
22:06
<sbalneav>
Plus, injure one man, tie up 3 more, slow down a whole group.
22:06
<loather-work>
yup, and once you hear the first mine go off, you know where your enemy is
22:06
<sbalneav>
All for something you can make for about $35 in bulk, and toss out the back of a low flying airplane.
22:06
<loather-work>
and can then descend upon them with flamethrowers or whatever
22:06
<sbalneav>
Despicable things.
22:06
<loather-work>
yeah, land mines are downright evil
22:07
i wonder if the geneva convention has anything in it about their use
22:08
<sbalneav>
Not the Geneva convention, AFAIK, but there was that big anti-landmine treaty that a bunch of countries signed a few years back.
22:09CaScAdE^FarAway has quit IRC
22:09
<loather-work>
oh good.
22:10
that old adage, "all's fair in love and war" doesn't really hold true
22:10
<sbalneav>
http://www.icbl.org/treaty
22:10
Agitated for, and spearheaded by, Canada, I will note here with a certain amount of pride.
22:11
<Ryan52>
I think I don't know how to merge with bzr.
22:11
if there are no conflicts, it looks like it doesn't merge the changes >.<
22:11
<loather-work>
hey, i like canada.
22:11
even moreso now that i see this
22:11
<sbalneav>
bzr merge lp:blahblah not working?
22:12
<Ryan52>
I think I'm just insane.
22:12
ok, ya, that was it.
22:12
<loather-work>
of course you are
22:14johnny has left #ltsp
22:36alekibango has joined #ltsp
22:41johnny has joined #ltsp
22:42johnny is now known as Guest79814
22:47Guest79814 is now known as johnny
22:54
<Ryan52>
oops. now the qt branches ldm can't start a greeter...
23:05
<sbalneav>
Ryan52: One more bug. Pushing 962. Thanks for ditching the unused status.
23:12
<Ryan52>
okay, thanks.
23:17
so, one thing that has been bothering me about ldm is that when you enter a wrong password you have to wait for the 30 second timeout, then ldm says "no repsonse from ssh server" (or something like that) and restarts...could we fix that somehow? If you have an idea as to how to fix it, I will do the coding and testing work...maybe a regex? "expect"ing "[.:]" instead of ":"? and, if it was a . then assuming it's the ssh server saying permission denied. st
23:24
<vagrantc>
that seems like a regression ...
23:24
it used to respond fairly quickly
23:25
<Ryan52>
ya, I know. but that was because now it's only asking for one chance to enter the password. if you failed all 3 it had to do the 30 second timeout.
23:25* Ryan52 wonders why it's only asking for one chance
23:27
<sbalneav>
That was something stgraber wanted, because of people who entered the wrong userid. :)
23:27
It should probably be configurable.
23:27
<Ryan52>
it should probably work right on failed passwords. :)
23:28
cause still if you enter the worng userid you don't want to sit there for 30 seconds.
23:28
that could easily freak poeple out...people get nervous when nothing is happening.
23:30
<sbalneav>
Problem is, we're limited by what ssh can do
23:31
ssh -l nonexistantuser host is gonna ask you 3 times regardless
23:32
It can't tell you that the user doesn't exist because that could be used for cracking.
23:34
<Ryan52>
there's always "ctrl+alt+backspace", anyway :)
23:34* Ryan52 will be content with that.
23:36
<vagrantc>
Ryan52: what did you ask me to test?
23:37Egyptian[Home] has quit IRC
23:38Egyptian[Home] has joined #ltsp
23:38
<Ryan52>
vagrantc, the multiple guest server thing. if you can't/don't want to, that fine...
23:38
<johnny>
sbalneav, wouldn't it be recommended to make it die after one shot?
23:38
that would make it behave better..
23:38alkisg has joined #ltsp
23:38
<Ryan52>
johnny: it does. but then ldm is still waiting for something to happen.
23:39
<johnny>
it won't if the sshd is set to allow 1 login attempt before starting over
23:39
<vagrantc>
Ryan52: ah yes.
23:39
Ryan52: well, i've got it up, i just need to test it :)
23:39
<sbalneav>
Let me have a look, one sec.
23:40
<johnny>
so.. the fedora experiment is working out well
23:40
just need a full release now.. :)
23:40* johnny <3 PackageKit
23:41
<vagrantc>
just when i'm ready to test ldm today, seems like Ryan52 and sbalneav have a bunch more patches
23:41
<johnny>
finally.. linux can do things that windows can.. in a way that can be standard across distros..
23:41
go go packagekit
23:42
<sbalneav>
Hmm
23:42
hold on, this should be fixable.
23:42
<johnny>
richard hughes is an amazing fellow..
23:42
<sbalneav>
uno momento, por favour
23:42
<johnny>
then again.. so is greg kroah-hartmann
23:44
<vagrantc>
Ryan52: initial test didn't go so well
23:46
Ryan52: actually, i had it misconfigured
23:48
Ryan52: works!
23:48
<Ryan52>
yay!
23:48
<sbalneav>
hm, lets see if this works.
23:48
<Ryan52>
so, it works to a point where freegeek has all of it's features verified working?
23:49
<vagrantc>
used LDM_GUEST_SERVER="2 3" and LDM_SERVER="1 2 3"
23:49
Ryan52: i think so.
23:50
<Ryan52>
yay!
23:50
<vagrantc>
Ryan52: haven't checked if it works with hostnames... but that's the only thing
23:50
<Ryan52>
I don't know why it wouldn't...
23:50
<vagrantc>
right
23:51
we do all that wacky tweaking of /etc/hosts at bootup
23:52
interesting thing to note: ssh connections made with -S don't actually need to have the right server ...
23:53
would work just as well to use "foo"
23:53
<sbalneav>
Hm
23:53
Well, one minor problem to solve, one sec
23:53
<vagrantc>
because the tunnel decides which server and what username it connects with
23:54japerry_cat has quit IRC
23:54
<vagrantc>
though it's probably good to put something in there that makes sense :)
23:55
i haven't yet gotten to tweaking gadi's xinitrc.d stuff
23:59
oh, gadi. introduced a debian-ism
23:59
run-parts --list ...