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


Channel log from 25 June 2008   (all times are UTC)

00:18cyberorg has quit IRC
00:18elisboa-1ff has joined #ltsp
00:34plamengr has joined #ltsp
00:38cyberorg has joined #ltsp
00:40elisboa-off has quit IRC
00:41japerry has quit IRC
00:45monteslu has joined #ltsp
00:54savetheWorld has quit IRC
00:58mistik1 has joined #ltsp
01:03subir has joined #ltsp
01:13deavid has joined #ltsp
01:17subir has quit IRC
01:27subir has joined #ltsp
01:30Pascal_2 has quit IRC
01:43Pascal_1 has joined #ltsp
01:54subir has quit IRC
01:55subir has joined #ltsp
01:57savetheWorld has joined #ltsp
02:06juanpaul has quit IRC
02:06subir_ has joined #ltsp
02:17subir has quit IRC
02:21vagrantc has quit IRC
02:33subir_ has quit IRC
02:34subir_ has joined #ltsp
02:54ccherrett has quit IRC
02:54ccherrett has joined #ltsp
03:18Rucal has joined #LTSP
03:19elisboa-off has joined #ltsp
03:23chrisjrob has joined #ltsp
03:24elisboa-2ff has joined #ltsp
03:29elisboa-2ff has quit IRC
03:29elisboa-off has quit IRC
03:29savetheWorld has quit IRC
03:29deavid has quit IRC
03:29elisboa-1ff has quit IRC
03:29jammcq has quit IRC
03:29primeministerp has quit IRC
03:29jbrett has quit IRC
03:29GiantPickle has quit IRC
03:29warren has quit IRC
03:29rjune has quit IRC
03:29loather has quit IRC
03:29popey has quit IRC
03:29plamengr has quit IRC
03:29elisboa-3ff has joined #ltsp
03:29elisboa-off has joined #ltsp
03:29savetheWorld has joined #ltsp
03:29deavid has joined #ltsp
03:29elisboa-1ff has joined #ltsp
03:29jammcq has joined #ltsp
03:29rjune has joined #ltsp
03:29loather has joined #ltsp
03:29jbrett has joined #ltsp
03:29warren has joined #ltsp
03:29GiantPickle has joined #ltsp
03:29primeministerp has joined #ltsp
03:29popey has joined #ltsp
03:33Nuba1 has joined #ltsp
03:37elisboa-1ff has quit IRC
03:38
<chrisjrob>
Hi All! I'm having a problem getting sane working under LTSP5 on Debian Etch...
03:38
firstly installing sane in the chroot gave me some errors...
03:38
libgphoto2-2: udev requires a mounted procfs, not started. failed!...
03:38
invoke-rc.d: initscript udev, action "reload" failed...
03:38
libsane: /sbin/MAKEDEV: warning: can't read /proc/devices...
03:39
predictably scanning isn't working. Can't find anything by googling... thoughts?
03:39
<rjune>
short answer is that mount -t proc /proc /proc needs to happen
03:40
I dont' know why it isn't happening
03:40
<ogra>
because it doesnt hapen magically in chroots ;)
03:40
you need to ount it after chrooting and unmount before you leave the chroot
03:41
*mount
03:41
<chrisjrob>
are these messages a problem?
03:41Q-FUNK has joined #ltsp
03:42elisboa-off has quit IRC
03:42
<ogra>
there should be more than just that
03:42
(more messages)
03:42
these should have told you if it failed or not
03:43
<chrisjrob>
not really, I have the full console from the install process and
03:43
it looked pretty clean other than that
03:43
but I do have errors when attempting to use the scanner
03:43
scanimage: unable to open /var/run/hplip/hpssd.port: No such file or directory: prnt/hpijs/hplip_api.c 94
03:44
scanimage: unable to connect hpssd socket 50002: Connection refused: prnt/hpijs/hplip_api.c 719
03:44
unsure whether they're related to the install messages or not.
03:45
<ogra>
well, i meant apt messages
03:45Nubae has quit IRC
03:46
<chrisjrob>
none from the console screens that I have ogra, but I will force a reinstall and
03:46
take another look.
03:46
should i be entering "mount -t proc /proc /proc" before apt-getting?
03:47
<rjune>
ogra, do you have to do PROC="Y" in lts.conf to make it happen?
03:49
<ogra>
rjune, lts.conf isnt involved in chroots on the server at all
03:49
chrisjrob, right and dont forget to unmount it before leaving the chroot
03:49
<rjune>
I missed that bit
03:49
<chrisjrob>
will do, thanks ogra, thanks rjune
03:50
<rjune>
I was thinking this was on a terminal
04:01Subhodip has joined #ltsp
04:17mikkel has joined #ltsp
04:20
<chrisjrob>
libsane is happy now, but libgphoto2-2 failed with...
04:20
udev requires a mounted sysfs, not started.
04:27Pascal_2 has joined #ltsp
04:28Pascal_1 has quit IRC
04:34Pascal_2 is now known as Pascal_1
04:36
<ogra>
chrisjrob, same as proc .... mount -t sysfs sys /sys
04:36
rarely needed but some packages look for devices during install
04:36
<chrisjrob>
thanks ogra
04:38
ogra should the first sys be /sys?
04:38
<ogra>
no
04:39
<chrisjrob>
thanks ogra
04:39
<ogra>
copy and paste should work
04:40
<chrisjrob>
worked perfectly ogra - now I'll need a quick client reboot to test!
04:41indradg_ has quit IRC
04:44chrisjrob has left #ltsp
04:52indradg has joined #ltsp
04:53subir_ has quit IRC
04:53chrisjrob has joined #ltsp
05:02subir has joined #ltsp
05:03captain_1agnus has joined #ltsp
05:07tjikkun_work has joined #ltsp
05:10captain_magnus has quit IRC
05:15basanta has joined #ltsp
05:24The_Code has joined #ltsp
05:39indradg has quit IRC
05:42The_Code_ has joined #ltsp
05:42The_Code has quit IRC
05:43The_Code_ is now known as The_Code
05:59ogra_cmpc has quit IRC
06:14ogra_cmpc has joined #ltsp
06:16alkisg has joined #ltsp
06:23The_Code has quit IRC
06:23basanta has quit IRC
06:25Subhodip has quit IRC
06:32indradg has joined #ltsp
06:36jammcq has quit IRC
06:46daya has joined #ltsp
06:47
<daya>
HI, in lenny tftp timeout occurs, whats the problem, its well in inetd.conf file,
06:47
<ogra>
can you tftp manually ?
06:52
<daya>
ogra, manually means tftp localhost or IP
06:52
<ogra>
localhost should suffice for a start
06:52
to see everything works
06:53
if that works its likely your dhcpd setup thats wrong
06:53
<daya>
ogra, didn't get you in 1st line
06:54
<ogra>
tftp localost should be enough to see if the tftp server works
06:59deavid has quit IRC
07:00
<daya>
ogra, but ps aux |grep tftp is now showing nothing, but tftp localhost gets work, showing tftp>
07:00
prompt
07:01
<ogra>
can you get a file ?
07:01
if that works all is fine
07:02
i.e.
07:02
get ltsp/i386/pxelinux.0
07:02
hit ctrl-d after that and check if pxelinux.0 lies in the dir you are in
07:04subir has quit IRC
07:05jammcq has joined #ltsp
07:06mhterres has joined #ltsp
07:10soneyka has joined #ltsp
07:15
<daya>
ogra, Transfer timeout
07:17
ogra, do I have to provide absolute path to /var/lib/tftp....
07:17
<ogra>
just type in what i wrote above (copy/paste)
07:17
get ltsp/i386/pxelinux.0
07:18
<daya>
yes, it do it, and it says Transfer Time out
07:18
<ogra>
fi that doesnt work something with tftp is wrong, if it does work its more likely dhcp
07:18
tftp is in /etc/inetd.conf ?
07:18
and /var/lib/tftpboot contains the needed files ?
07:19
<daya>
why ps aux |grep tftp is not showing any process status
07:20
<ogra>
because inetd will only fire it up during actual transfer
07:20
<daya>
oh,
07:20
<cyberorg>
ogra, does this still work? http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspInfo
07:20
<ogra>
htop or top might show it for a second or two
07:20
luckily not :)
07:20
<daya>
but Transfer Tiime out ,
07:20
<ogra>
that was hilariously insecure
07:20
<cyberorg>
cool :)
07:21bartonello has joined #ltsp
07:22
<ogra>
cyberorg, https://wiki.ubuntu.com/LTSPClientShutdownReboot
07:22
something like that hould work with new ltsp
07:23
but even that is sily and insecure
07:23
<cyberorg>
ogra, yeah, italc seems like a best option
07:23
<ogra>
the proper solution is to implement the localapps spec, then such stuff can hook into it
07:23elisboa-3ff has quit IRC
07:24
<ogra>
i hope we can attack it at the hackfest next month
07:25chrisjrob has left #ltsp
07:26
<ogra>
hmm, that reminds me
07:26
<cyberorg>
ogra, would you be around tomorrow same time? we'd be having our weekly meeting, we'd make it here as no one joins us in our soc channel :)
07:26* ogra makes a note to enable xrexecd in the next ltsp 5.1.10 upload
07:27
<ogra>
cyberorg, cant tell yet, i' massvely busy with subnotebook and classmate stuff
07:27
<cyberorg>
now we are more on lts.conf territory
07:28
<daya>
ogra, I can't get what is distrubing tftp to start,
07:28
<bartonello>
hello. Could anyone help me to connect my usb printer to ltsp thin client? I`ve added " PRINTER_0_DEVICE=/dev/usblp0" and "PRINTER_0_TYPE=U" to "[default]" section of lts.conf file at tftpboot directory. But after booting thin client I can`t see any /dev/usblp0 device. May be I`ve forgot something? Sorry for my English.
07:28
<ogra>
daduke, you didnt answer my questions :)
07:28
err
07:28
daya, ^^^^
07:28
<daya>
ogra, :)
07:29
<ogra>
is tftp in /etc/inetd.conf ? and does /var/lib/tftpboot contain the needed files ?
07:29
(and is inetd even running at all )
07:30
<daya>
ogra, yes, all is to default setting
07:31
ogra, inetutils-inetd is running
07:31
<ogra>
and /var/lib/tftpboot has the ltsp subdir with kernels and pxe stuff etc ?
07:31
<daya>
yes of course
07:32
<ogra>
well, the error would be file not found anyway if they were missing
07:32
<daya>
ogra, should xinetd work
07:33* ogra wasnt aware that debian ltsp switched to inetutils-inetd
07:33
<ogra>
i dont think vagrant included the config for xinetd
07:33
there are files for it in the source from redhat though
07:35
<daya>
ogra, inetutils-inetd isn't it the required pkgs for inetd
07:35juanpaul has joined #ltsp
07:35
<ogra>
openbsd-inetd is what we used to use
07:36
not sure if vagrant changed that, in ubuntu its still the default inetd in the ltsp-server dependencies
07:37
<daya>
ok i will try for openbds-inetd, though it was not asked while I am builing ltsp
07:37
<ogra>
no, its a package dependency
07:38
it would have been installed with the ltsp-server package ... if you dont have any inetd already
07:38
if you installed inetutils-inetd before that ltsp-server assumes you know what you do and doesnt try to install another inetd
07:42
<daya>
ogra, then I should not remove inetutils-inetd
07:43
ogra, But I already did it,
07:43
<alkisg>
ogra, about ipconfig, I've send a patch (accepted in last klibc version) which fixes most of the "2 dhcp server" cases without changing ltsp_nbd. There was a bug, DHCPOFFER could be mistaken for DHCPACK.
07:44
<ogra>
alkisg, bah, i just merged the ltsp_nbd fixes :)
07:44
<daya>
with opendsd , file is received successfully in localhost
07:44
<ogra>
daya, so try booting a client
07:44
<alkisg>
ogra, heh! OK, if/when you have the time, please use the latest klibc.
07:45
<ogra>
alkisg, but no prob to drop them if the klibc fix enters ubuntu :)
07:45
<alkisg>
ogra, what fixes? binary transfer? any link?
07:46
<ogra>
no, just the "additional port for a second dhcpd" fix
07:46
<alkisg>
oh, I thought that was already there, about port 1068...
07:46
I'll look it up.
07:46
<daya>
ogra, it gets worked , client boots successfully
07:47
ogra, may be there is problem in inetutils-inetd
07:47
in lenny
07:47
<ogra>
http://bazaar.launchpad.net/~ogra/cmpc/cmpc-gen1-installer
07:47
gah, wrong paste
07:47
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/revision/737
07:47
alkisg, ^^^ that one
07:48
it was in ubuntu hardy, but i didnt come around to merge it upstream yet
07:48
daya, yes, likely ... thats why the package depends on openbsd-inetd by default ;)
07:51
<daya>
ogra, ok thanks for resolving my prob.
07:51
<ogra>
youre welcome :)
07:54
<alkisg>
ogra, I'm probably missing something, but this code for DHCPPORT has been in ltsp_nbd since many months now... Anyway, nevermind! :)
07:54Q-FUNK has quit IRC
07:56
<alkisg>
OK, I've just read your previous line "it was in ubuntu hardy, but i didnt come around to merge it upstream yet". I'm just plain stupid. :P
07:56
<ogra>
alkisg, in ubuntu, yes :) but i never pushed it upstream (the link i gave was the upstream branch)
07:58
<alkisg>
ogra, OK, you don't have to drop them even if the new ipconfig gets in ubuntu, it solves a different problem... Some people may actually need a different port.
07:58
I've implemented a dhcp-server selection phase in ipconfig, but I don't think hpa will ever accept it upstream.
07:58juanpaul has quit IRC
07:58
<ogra>
yeah, i rememberd that afte i wrote that sentence
08:03Rucal has quit IRC
08:13
<daya>
ogra, in chrooted evn, /opt/ltsp/i386 i set the passwd for root, but in client it doesn't let me to login in tty1
08:13juanpaul has joined #ltsp
08:13
<ogra>
daya, you also need to unlock it
08:13
<daya>
ogra, unlock thru passwd file
08:13
<ogra>
(at least if login attempts tell you its lockaed)
08:13
no, i mean deliberately with the passwd command
08:14
<daya>
ogra, I have set the passwd using passwd command, but says account expired
08:14
<ogra>
right
08:14
unlock it
08:15
<daya>
how to unlock, :D, I have issued passwd and kept the password for root, but it doesn't solve
08:17
<ogra>
man passwd ;)
08:17
cyberorg, you default to CONFIGURE_X=False in suse ?
08:17
how do you get the keyboard set up ?
08:18
<cyberorg>
ogra, no
08:18
<ogra>
ah, the recent mail to ltsp-discuss seemed like that
08:18
<cyberorg>
using configure-x.sh of course
08:18Pascal_1 has quit IRC
08:20slidesinger has joined #ltsp
08:22rafaelcardoso has joined #ltsp
08:22
<ogra>
cyberorg, ltspfs should have an example script for putting links on KDE desktops
08:22
iirc vagrant pushed that upstream
08:23
<rafaelcardoso>
wilisystem itś alive?
08:23
<cyberorg>
ogra, ah, cool, will look those up, would those work for kde4 too?
08:23
<ogra>
well, its an evil hack that just does ln /media/whatever to ~/Desktop
08:24
so it should work everywhere
08:24
http://bazaar.launchpad.net/~ltsp-upstream/ltspfs/ltspfs-trunk/annotate/vagrant%40freegeek.org-20080617034718-zw31rwt60mh223h2?file_id=kdedesktopicons-20080302225737-q6d8ttvp6y7nfi4a-2
08:24
<cyberorg>
ogra, kde4 desktop is plasma, it is like mystery to me, don't understand much of it :)
08:24
<ogra>
tahst the snippet
08:25
<rafaelcardoso>
hi all wilisystem.com works?
08:27
<ogra>
cyberorg, oh, i see it improved, it doesnt blindly link anymore but create a .desktop file with link url
08:33daya has quit IRC
08:35
<cyberorg>
that looks good
08:37alkisg has quit IRC
08:38
<ogra>
still evil to put files in user dirs
08:38
but less than before at least
08:39likuidkewl has joined #ltsp
08:54mikkel has quit IRC
09:03rafaelcardoso has quit IRC
09:14alkisg has joined #ltsp
09:23gdi2k has joined #ltsp
09:25
<gdi2k>
Greetings. I've bought myself an Intel Atom based motherboard for experimentation with LTSP - my hope is that it will be quite a cheap but capable thin client. However, I can't get it to boot. I'm using a default LTSP 5 install on Hardy. I have a Via C7 booting fine off the same system. It gets to the Ubuntu splash screen, but then halts. Any ideas?
09:27bartonello has quit IRC
09:28
<warren>
what kind of video card does the atom have?
09:28
just curious
09:28
gdi2k: very few of us have even seen an atom
09:30
<gdi2k>
yes, they're still quite new
09:30
I'd be happy to run whatever on it to help you learn more about it...
09:31
<warren>
can you do lspci?
09:31
what video is it?
09:32
<gdi2k>
I would have to boot off CD to get that far, will take some time. I have this model: http://www.tranquilpc-shop.co.uk/acatalog/Motherboards.html which has an Intel GMA 950 graphics chip
09:38
<ogra>
warren, intel usualy
09:38
<warren>
ogra: I've seen atom with "powervr" video which incredibly sucks
09:38
gdi2k: 950 is pretty good
09:38
<ogra>
warren, though you could put in others i doubt intel will sell boards without onboard
09:39
ATOM is massively impressing
09:39
<warren>
I can't wait to buy a mini notebook
09:39
<ogra>
my classmate just flies and can nearly cope with my core duo lappie
09:39
<jcastro>
have you tried it as thin client yet?
09:39
<ogra>
within the next 4 weeks i'll releae the classmate imge publically
09:40
<gdi2k>
I'm trying! It halts on boot, during when the splash screen shows up
09:40
<ogra>
jcastro, not yet, it would needs some fiddling with ltsp (planned)
09:40
<jcastro>
an atom-based thin client would be so wicked
09:40
<gdi2k>
what is the issue with ltsp - I thought it was just a normal i386-based machine that would have no issues with ltsp?
09:40
<ogra>
during intrepid i'll add lpia support to the scripts so ATOM should waork then
09:40
*work
09:41
<warren>
lpia?
09:41
<jcastro>
low power intel arch
09:41
<ogra>
ATOM specific arch
09:42
we have that in ubuntu for mobile and netbooks
09:42
<gdi2k>
so it's not i386 compat?
09:42
<ogra>
it makes use of the power saving features
09:42
<warren>
Google says "Libertarian Party of Iowa"
09:42
<gdi2k>
same thing, right?
09:42RoninBaka has joined #ltsp
09:42
<ogra>
(my classmate has 4h battery life on the celeron and with the same battery 6.5 on the atom with lpia)
09:43
gdi2k, i386 should essentially work for the basics
09:43
<warren>
so you need a special kernel built for it?
09:43
<ogra>
we have a special toolchain and kernel for it, yes
09:44
i dont think you *need* it but to optimize for that arch it makes sense to have it
09:44
<gdi2k>
ok that's good news. how do I go about diagnosing my issue with booting?
09:44
<jcastro>
yeah I think you can run normal i386 kernels on it but you don't get the powersaving features
09:44
<warren>
are there any retail portables with intel atom yet?
09:44
<ogra>
is that ubuntu ?
09:44
<RoninBaka>
is there a way to start a ltsp session from within an existing windows bootup?
09:44
<gdi2k>
yes, hardy
09:45
<ogra>
gdi2k, remove quiet and splash from /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
09:45
that gives you a verbose textboot
09:45
warren, there will be soon ...
09:47
<warren>
gdi2k: you bought a motherboard with atom?
09:47
<gdi2k>
ogra: ok, just rebooting.
09:48
warren: yes
09:48
hmm, it has an "abnormal exit" when running modprobe
09:48
<ogra>
ah
09:48
<warren>
http://www.newegg.com/Product/Product.aspx?Item=N82E16813121342&Tpk=intel%2batom
09:48
<ogra>
well, then you likely need the lpia build
09:48
<gdi2k>
I'll take a photo so you can see what's going on...
09:49
<warren>
this page says you have to disable the ethernet in the BIOS in order to install Ubuntu 8.04
09:49
<ogra>
i was suspecting the drm of the i915 module before :)
09:50
warren, the relatek breakage was/is fixed in 8.04.1
09:50
<warren>
http://www.tranquilpc-shop.co.uk/acatalog/T7Atom.html#a72 looks like somebody built an atom-based thin client
09:50
<ogra>
well, all the "nettops" actually re thin clients with disks
09:50
*are
09:51
and i guess not even expensive if the cost for the disk can be cut
09:51mikkel has joined #ltsp
09:52
<warren>
this page doesn't clarify if the unit is passive cooled
09:52
or it runs a fan
09:53
<ogra>
my cmpc has a fan
09:53
<warren>
oh, nm,
09:53
<gdi2k>
mine is not - the CPU requires no active cooling, but the northbridge does...
09:53
<warren>
at the top it says fanless
09:53
<ogra_cmpc>
ogra@classmate:~$ cat /proc/acpi/thermal_zone/THRM/trip_points
09:53
critical (S5): 104 C
09:53
and pretty intresting trip point values :)
09:55
<ogra>
warren, we just released our mobile edition that runs on lpia ... and will sonn have netbook images as well
09:57
<warren>
I wonder how its CPU performance compares
09:57
<gdi2k>
ok, here's the issue: http://farm4.static.flickr.com/3035/2610010663_ac3452f277_o_d.jpg
09:58
since then, it's spat out another couple of lines about not being able configure eth0, followed by a kernel panic
09:58
<ogra>
warren, well, i can compile th eubuntu kernel package within 2h on the cmpc with 512M
09:58
it takes a bit less on my lappie with core duo and 3G though :)
09:59
<warren>
gdi2k: did you try disabling the ethernet in the bios?
09:59
<ogra>
gdi2k, can you scroll up
09:59
warren, bad idea if you netboot :)
09:59
it looks like the realtek error
09:59
the top line of the oops could tell
10:00
anyway
10:00
<gdi2k>
will disable bios network booting. can't scroll up...
10:00
<ogra>
gdi2k, enable hardy-updates in the sources.list in the client chroot
10:00
and install linux-image-2.6.24-19
10:00
that has a fix
10:01
<gdi2k>
ok
10:01
<ogra>
then run ltsp-update-kernels on the server to get the kernel into tftpboot
10:01
if the realtek breakage is the only prob it will work then
10:02* ogra ponders if he should add support for arm thin clients as well ... we'll have it supported in intrepid
10:03
<ogra>
does anyone know anybody who owns arm terminals ?
10:03mccann has joined #ltsp
10:04
<gdi2k>
just normal updates, not pre-release, right?
10:04
<ogra>
yes
10:05
easiest is to do it in the gui on the server and just copy sources.list over then
10:11
<gdi2k>
ok, got it. downloading on my incredibly slow and expensive connection...
10:12
(and censored)
10:13indradg_ has joined #ltsp
10:17RiXtEr has joined #ltsp
10:17indradg__ has joined #ltsp
10:21
<gdi2k>
success! thanks for that! and it boots very fast too!
10:22
<ogra>
good :)
10:23
warren, hey, looks like i'll be in lexington before portland
10:23
<warren>
lexington?
10:24
<ogra>
near boston
10:24
or part of boston ?
10:24Pascal_1 has joined #ltsp
10:24
<warren>
oh, what are you doing in boston area?
10:24
<ogra>
our mobile solutions team has the office there
10:25
i was actually planning to do my distro sprint in london ... but was just told it makes likely more sense to go there since i work more and more with mobile and netbook stuff
10:26* ogra waits for final word from maagement
10:27
<ogra>
(we do a sprint where all devs meet in person for a week in the middle of a release cycle)
10:27Pascal_1 has quit IRC
10:27
<ogra>
since we all work from home otherwise
10:27
<gdi2k>
sounds like I should apply! ;-)
10:28
<ogra>
http://webapps.ubuntu.com/employment/
10:28
<gdi2k>
the atom runs nicely as a thin client by the way!
10:28
<ogra>
if you got the skills :)
10:28* warren highly considers buying this T7 Atom thin clienet
10:28
<warren>
it sounds really cool
10:28
<gdi2k>
sadly I don't think I do :(
10:29likuidkewl has left #ltsp
10:31
<gdi2k>
so how are things looking on the local apps front for intrepid?
10:31
<warren>
hopefully we'll have that working in LTSP upstream at this upcoming LTSP hackfest...
10:31
<ogra>
we'll be able to tell you after the hackfest i guess
10:31
warren, do you build xrececd already ?
10:31
we should all enable it in advance
10:32
<warren>
I have no idea what it is
10:32
<gdi2k>
excellent - I think the newer generation low power CPUs are quite capable of offloading the servers...
10:32
<ogra>
so we only need to care for the glue
10:32
warren, its in the ltsp source
10:32
<warren>
xrececd wont need further improvements?
10:32
<ogra>
it likely will
10:32
but the binary should already do the basics
10:32
<warren>
oh i see it
10:33
<ogra>
(listen on the rootwin for the execution commands)
10:33
<warren>
the source dir has iptables rules
10:33
is that for the client or server?
10:33
<ogra>
iptables ?
10:33* ogra never saw that
10:33
<warren>
ltsp-trunk/client/xrexecd/iptables.sh
10:33
<ogra>
uuuh
10:34indradg has quit IRC
10:34
<ogra>
i dont think we need these
10:34
<warren>
ogra: do you already have a path to install xrexecd?
10:35Subhodip has joined #ltsp
10:35indradg has joined #ltsp
10:35
<ogra>
i'd put it in /usr/sbin
10:35
<warren>
a quick glance at the code, makes me wonder how secure this is.
10:35exodos has joined #ltsp
10:35staffencasa has joined #ltsp
10:36
<ogra>
well, it surely needs improvement
10:36elisboa has joined #ltsp
10:36
<ogra>
but its totally tied to X so however secure you consider xauth ....
10:36
<cyberorg>
i've got it in /usr/bin/xrexecd
10:36
<warren>
my client already has xrexecd installed
10:37
<ogra>
oh
10:37
<warren>
in /usr/bin
10:37
<ogra>
i dont install it yet
10:37* ogra will do so for the intrpid packae
10:38
<warren>
bfeore implementing the keyboard layout thing I'm trying to make it run the session correctly
10:38
little confused by the ldm code
10:38
<ogra>
we need a second parameter for that
10:38jerhum has joined #ltsp
10:38
<ogra>
ldminfo.session now carries the Xsession path
10:39
<jerhum>
hello all
10:39
<warren>
and ldm was replacing the content of ldminfo.session?
10:39
<ogra>
no
10:39
if the var is set (from lts.conf) that value overrides
10:39
else it wil use the distro specific path
10:39
<warren>
we might not be talking about the same thing
10:40
<ogra>
(there is a huge case statement for the different paths of Xsession)
10:40
<warren>
I mean, if you select KDE, then run /path/to/Xsession startkde
10:40
<ogra>
right
10:40
<jerhum>
where find screenscript for debian etch for rdesktop ?
10:40
<ogra>
currently ldminfo.session is used for that but a custom session you select will not be executed through Xsession atm
10:41
so we need two parameters or append to ldminfo.session to get that going
10:41
<warren>
the latter is fine to me
10:41
<ogra>
but the code needs some massive changes
10:41
<warren>
why is this particular thing massive?
10:41
I'm thinking about adding a bunch of comments to ldm though
10:41indradg__ has quit IRC
10:41
<ogra>
because currently ldminfo.session is hogged by Xsession
10:41tjikkun_work has quit IRC
10:41
<warren>
so someone from GNOME reading the source cna more easily understand what is going on
10:42indradg_ has quit IRC
10:42
<ogra>
hmm
10:42* ogra never had probs to understand it :)
10:42
<ogra>
even the parts scott wrote :)
10:43
but we need to rework the part that deals with ldminfo.session in any case for that
10:43
<warren>
we don't actually need another lts.conf option for this
10:43
right?
10:43
<ogra>
no
10:43
but another variable in the code
10:44
one that carries the distro path to Xsession and one that has the actual session call
10:44
<warren>
the word "session" is really overloaded in here
10:44
<ogra>
since the distro paths get set dynamically on login
10:44
heh, yes
10:44
<warren>
referring to three different things
10:45
LDM_SESSION is an optional thing to point at the Xsession script right?
10:45
<ogra>
the current ldminfo.session should become ldminfo.xsessionpath
10:45
<warren>
I'm considering defining that by default in our lts.conf so any ldm client (especially flashed on a disk) can Just Work
10:45
<ogra>
or tostartkde or gnome-session or so
10:45
well, that would break as soon as you connect to any non redhat server from the client
10:46
ldminfo.xsessionpath needs to be determined dynamically at login
10:46
<warren>
we talked about this earlier
10:46
I want arbiitrary OS thin client with ldm to be able to login to any OS server.
10:46
<ogra>
while ldminfo.session shuld just call the actual startkde or gnome-session
10:47
and by default should just be set to "default"
10:47
<warren>
so the server has to somehow tell the client
10:47
either from lts.conf or ldminfod
10:47
<ogra>
the actual command will then be merged from ldminfo.xsessionpath and ldminfo.session
10:47
no
10:47
look at the code
10:47
<warren>
sorry, i'm talking about a different problem
10:47
<ogra>
it determines the Xsession path uring login
10:48
*during
10:48
<warren>
huh?
10:48
no it doesn't
10:48
the client knows a Xsession path from the local filesystem itself
10:48
which may not match the server.
10:49
<ogra>
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ldm-trunk/annotate/vagrant%40freegeek.org-20080530224058-kjzuqmfbe3sih3e9?file_id=ldm.c-20070611142051-jkh2vwmsp70sequl-32
10:49
line 450 to 470
10:49
it tries access() on the different possible paths during login
10:49
<warren>
453 821 if (*ldminfo.session == '\0') {
10:49
454 if(!access("/etc/X11/xinit/Xsession", X_OK)) {
10:49
<ogra>
right
10:50
<warren>
ogra: yes, but that's entirely on the filesystem of the client
10:50
<ogra>
no
10:50
<warren>
eh?
10:50
<ogra>
thats after the first ssh conn is established
10:50
<warren>
you sure?
10:50
<ogra>
yes
10:50
<warren>
how is it using the access() function across the network?
10:50
this is running in ldm itself
10:50
<ogra>
at least it should be ... if not we need to fix it that way
10:51
<warren>
it is blindly running /path/to/Xsession over ssh without knowing if it is actually there.
10:51
<ogra>
i want to be able to select this redhat server from the serverlist in my ldm and just connect
10:51
<warren>
I agree.
10:51
currently ldm can't do that
10:51
unless you set LDM_SESSION in lts.conf
10:51
<ogra>
so it needs to run the access() on the server before the session gets started but after the first ssh connection is up
10:52
well, then lets fix it the right way :)
10:52
<warren>
We had a flamewar about this on the list before
10:52
It seems most logical that ldminfod would tell ldm where Xsession is located.
10:52* ogra wouldnt call that flamewar
10:52
<ogra>
but then you *need* ldminfod
10:52
<warren>
because then it wouldn't need yet another network query
10:52
without ldminfod doing it, then you need lts.conf
10:52* ogra doesnt like the idea to rely on ldminfod
10:52
<warren>
without either, then you need to query the server
10:53
<ogra>
nor to use lts.conf by default
10:53
right
10:53
but we dont need ldminfo in that scheme
10:53
since we have an ssh tunnel up already and can take a look on the server
10:53
<warren>
ok, so s/ldminfo.session/ldminfo.xsessionpath/
10:54
then add a new one for the GUI chosen session
10:54juanpaul has quit IRC
10:54
<warren>
I wonder how much that'll slow things down.
10:54rjune_ has quit IRC
10:55
<ogra>
i doubt its much
10:55
<warren>
what was the thing from last week that we were discussing
10:56
to also query over the network
10:56
<ogra>
we shuld probably have a script that rearranges the code dynamically for the building distro to be first in the list
10:56
<warren>
if you're having it query over the network
10:56
then having "first in the list" doesn't matter
10:56
<ogra>
you have an ssh tunel up already
10:57
<warren>
ogra: if that's the case, then perhaps it should be querying everything and we eliminate ldminfod?
10:57
<ogra>
i cant imagine it takes much time to call access() over it and get the return value
10:57juanpaul has joined #ltsp
10:57wwx has quit IRC
10:57
<warren>
OTOH, can't do that because you aren't logged in yet..
10:57
<ogra>
you are ;)
10:57
ldm uses two ssh tunnels
10:58
the first one is non graphicl
10:58RiXtEr has quit IRC
10:58
<warren>
no no
10:58
<ogra>
the seond one attaches to it and cals the actual session
10:58
<warren>
I mean, we could think about getting rid of ldminfod entirely
10:58
but we can't
10:58
<ogra>
no
10:59
well, we can but then it would always use default lang and session
10:59
<cyberorg>
if we have ldminfod for 10 things why cant we have 11th in it too?
10:59
<ogra>
whic is the good thing on ldm
10:59
it currently can connect to $random_ssh_server
10:59
because ldminfo is optional
11:00
<warren>
we're a bit inconsistent here
11:00
<ogra>
and i wouldnt like to see that going away
11:00* warren food
11:00
<ogra>
well, that behavior is consistent since the first day of ldm
11:00
and was very much in my focus when i wrote ldm
11:01
it needs to be able to connect to any ssh server without hving ldminfo
11:02
cyberorg, we have it, but we're still operational without it
11:02
<cyberorg>
what does nautilus use to find all sftp servers in the network, can we use that?
11:02
<ogra>
i dont mind adding another line to the ldminfo output but ldm needs to operate still if its not there
11:03
with the servers defaults
11:05
<cyberorg>
got that, i dont mind current if(!access("/etc/X11/xdm/Xsession", X_OK)
11:06
<ogra>
well, tats local
11:06
<cyberorg>
it shouldn't take much time doing that
11:06
<ogra>
warren is worried that it slows down things if the same thing goes through the ssh tunnel
11:06
right, i agree, i guess it will take less than a second
11:07
far less
11:07
<cyberorg>
at least it gives flexibility to connect to any ltsp server
11:07
<ogra>
even if we'd do it thrugh an rc.d script it wouldnt really delay
11:08
and we probably should test that since its easier to add more distros to script code than to the C code
11:08
keeping the determination completely out of ldm here and just seed a variable with the Xsession path
11:09
<cyberorg>
variable that can take multiple values, in case you have debian as well as gentoo server would be nice
11:09jerhum_ has joined #ltsp
11:10
<ogra>
well, you will only connect to one with one session
11:10
<cyberorg>
i agree this should be moved out of ldm.c and create variable in one of the scripts
11:10
<ogra>
it only nees to hold the path
11:11
and if it is an rc.d script the session is actually already established at that point, there can only be one value
11:11
i'd like to hear vagrantc about it though ...
11:11K_O-Gnom has joined #ltsp
11:11
<ogra>
he is always concerned bout filesystem calls
11:12
<cyberorg>
in that case we can just hard code the variable during package build time
11:12
<ogra>
no
11:12
you need to be able to connect to different servers
11:12
<cyberorg>
right
11:12
<ogra>
yu cant hardcode that value
11:12
but at the point where the variable gets set, the connection is already up
11:12Ryan52 has joined #ltsp
11:13
<ogra>
so you can only get one value (the one from the server you are connected to)
11:13alkisg has quit IRC
11:14alkisg has joined #ltsp
11:14|Ryan52 has quit IRC
11:14
<ogra>
hmm, actually we can do it all in script and have a proper LDM_SESSION variable filled in the end that ldm can just use
11:15
merged from Xsession path and session command
11:16japerry has joined #ltsp
11:22jerhum has quit IRC
11:24
<warren>
ogra: one thing to query may not be bad, but this isn't the first thing we proposed to query
11:28jerhum_ has quit IRC
11:30
<ogra>
warren, what else do we query apart from teh servers xauth atm ?
11:30
<warren>
are we querying? I thought we're setting.
11:31
<ogra>
oh, right we push
11:32
well, but i guess just having an rc.d script that pulls the path and merging the output wiht LDM_SESSION (or using "default" if thats not set) shouldnt really slow us down
11:32
and in such a script you can actually reorder the check paths for the distro you build on
11:33
<warren>
well
11:33Nubae has joined #ltsp
11:33
<ogra>
so your own path will always be checked first and return fast
11:33
<warren>
rc.d scripts are run in sequence and each blocks
11:33wwx has joined #ltsp
11:33
<ogra>
as soon as its a different distro it will take a millisecond longer
11:33
<warren>
the benefit of ldminfod was the server gives you all the output at once
11:34
<ogra>
sure, as i said feel free to add that info to ldminfod additionally ... but i want a dynamic way that works without ldminfo as well
11:34
so that rc.d script could only fire if there is no info
11:34
<warren>
You don't see it as bad to have several inconsistent ways of doing this?
11:35
<ogra>
whats inconsitent about that ?
11:35
<warren>
"no info" means we remove the hard coded one in ldm now
11:35
<ogra>
if ldminfo is there we take that ... if not we fal back to probing
11:35
<warren>
I suppose that doesn't break anything if rc.d queries it...
11:35
<ogra>
no
11:35
not remove .... we move :)
11:35
<warren>
ok, this isn't a bad idea.
11:36
ldminfod makes it fast by default, but probing is the fallback
11:36
<ogra>
move it out of C and make ist an rc.d shellscript
11:36
*it
11:36
<warren>
could you write the proposal to the list?
11:36
<ogra>
sure
11:36
<warren>
cool
11:36
I like this particular plan.
11:36exodos has quit IRC
11:37
<ogra>
but a bit later today i'm currently massively busy getting my flights sorted ...
11:37
<warren>
older ldm doesn't break (because it has a hard coded value)
11:37
<ogra>
my sprint ends on 19th ... hackfest starts at 24th
11:37
not sure what i should do with that gap
11:37
seems silly to fly from boston to germany for 3 days
11:38alkisg has quit IRC
11:38* ogra makes note to notify davidz so we can go out for a beer
11:39
<ogra>
we found we make pretty good drink partners in prague :)
11:42
warren, btw ... small atom notebooks, have you seen that ? http://people.ubuntu.com/~ogra/cmpc-netbookdesktop.png
11:42
<warren>
no but it looks nice.
11:42
except for the brownness
11:42
=)
11:42
<ogra>
all clutter based
11:42
so it needs GL support sadly
11:42
but has nifty animations and the like
11:42
<warren>
cool
11:42
<cyberorg>
saw these? http://www.msimobile.com/DetailPage.aspx?model=Wind_NB_Linux
11:43
<ogra>
this is what i'll be working on in the future
11:43
<warren>
does it suspend fine?
11:43
<ogra>
building netbook OSes
11:43
yep
11:43
it needs a hack to keep usb persistence for the SDD though
11:43
(kernel patch)
11:44
<warren>
80GB hard drive
11:44
<ogra>
so i cant use the ubuntu mainline kernel sadly ... thats the most painful part
11:45
the HDD will eat your battery
11:46
Operating System Linux Novell
11:46
heh
11:47
they dont specify battery life though ... the classmate with SSD USB disk survives 6.5h without me having added any optimization yet
11:47
i bet i can get it towards 7-7.5h with a bit fiddling
11:48
<warren>
why ssd usb/
11:48
why USB?
11:49
<ogra>
sily decision ...
11:49
<warren>
USB adds extra power consumption and all sorts of issues with linux kernel
11:49
<ogra>
but seems cheaper
11:49
<warren>
oh
11:49Nuba1 has quit IRC
11:49
<ogra>
intel asked me for input and they actually built one device with ATA SSD for me
11:50
but then changed back to USB, i havent gotten any answer yet why they decided that but i guess its a matter of money
11:50
(the classmate has also no inte wlan for the same reason)
11:50
*intel
11:50
<warren>
what kind of wlan?
11:50
<ogra>
realtek
11:51
USB indeed :P
11:51
<warren>
what driver?
11:51
wow
11:51
<ogra>
rt74
11:51
err
11:51
73
11:51
<warren>
open source?
11:51
<ogra>
yep
11:51
<warren>
oh
11:51
well, the USB part sucks
11:51
<ogra>
you should have it in your kernel
11:51
yeah, a bit
11:51
<warren>
ok, s/ldminfo.session/ldminfo.xsessionpath/
11:52
what should we call the new session variable
11:52
<ogra>
but i have overcome the bitterness here already ... the ATOM makes it really rock ... on the celeron version it sucked all your IO
11:52
<warren>
sessiondesktop?
11:52
<ogra>
well, i think we actually dont need that
11:52
we can merge it in the rc.d script
11:53
<warren>
how long will it take to implement that?
11:53
<ogra>
either from the ldminfo return value or from the query
11:53
i can do it on my flight to boston :)
11:53
<warren>
ok... I'll just focus on the keyboard layout for now then
11:53
how do I revert the entire tree in bzr...
11:55
<ogra>
bzr revert without options ?
11:55
should work if you havent committed
11:55
<warren>
ohi see
11:56
<ogra>
if you have committed, you need to uncommit first
11:57
<warren>
hmm, we aren't using create_xauth()
11:57
<ogra>
no, thast done in the rc.d script iirc
11:58
<warren>
I'll add a setup_keyboard_layout() and run it after launch_x
11:58
ogra: IIRC, half (creation of the key) is done in screen scripts, the other half (setting the server) is rc.d
11:59
<ogra>
right
11:59
but nothing in the C code anymore
12:00* ogra prefers scripts where possible and where no slowdown is involved since people can easier contribute
12:00
<ogra>
i was really sad to drop the python ldm ... we had far more people sending patches etc with it
12:02captain_1agnus has quit IRC
12:02twinprism has quit IRC
12:02* ogra sighs and returns to http://merges.ubuntu.com/main.html to get these last 52 packages touched :/
12:02captain_1agnus has joined #ltsp
12:02twinprism has joined #ltsp
12:02Ryan52 has quit IRC
12:02|Ryan52 has joined #ltsp
12:03japerry has quit IRC
12:03elisboa has quit IRC
12:03mhterres has quit IRC
12:03ogra_cmpc has quit IRC
12:03GiantPickle has quit IRC
12:03rjune has quit IRC
12:03loather has quit IRC
12:03warren has quit IRC
12:03primeministerp has quit IRC
12:03jbrett has quit IRC
12:03popey has quit IRC
12:03savetheWorld has quit IRC
12:03jammcq has quit IRC
12:03gentgeen__ has quit IRC
12:03MacIver has quit IRC
12:03Lns has quit IRC
12:03F-GT has quit IRC
12:03mccann has quit IRC
12:03Patina has quit IRC
12:03loather-work has quit IRC
12:03dberkholz has quit IRC
12:03ogra has quit IRC
12:03twinprism has quit IRC
12:03captain_1agnus has quit IRC
12:03|Ryan52 has quit IRC
12:03mistik1 has quit IRC
12:03sutula has quit IRC
12:03iMacGyver has quit IRC
12:03Lumiere has quit IRC
12:03artista_frustrad has quit IRC
12:03cyberorg has quit IRC
12:03klausade has quit IRC
12:03gu has quit IRC
12:03xachen has quit IRC
12:03gonzaloaf_work has quit IRC
12:03jcastro has quit IRC
12:03|Paradox| has quit IRC
12:03indradg has quit IRC
12:03topslakr has quit IRC
12:03irule has quit IRC
12:03fsgaston has quit IRC
12:03Egyptian[Home] has quit IRC
12:03wwx has quit IRC
12:03mikkel has quit IRC
12:03vpro has quit IRC
12:03Subhodip has quit IRC
12:03K_O-Gnom has quit IRC
12:03slidesinger has quit IRC
12:03RoninBaka has quit IRC
12:03ccherrett has quit IRC
12:03ufo_ has quit IRC
12:03yanu has quit IRC
12:03sep has quit IRC
12:03topslakr_ has quit IRC
12:03laga has quit IRC
12:03moquist has quit IRC
12:03leio has quit IRC
12:03stgraber has quit IRC
12:03staffencasa has quit IRC
12:03daduke has quit IRC
12:03alekibango has quit IRC
12:03|Ryan52 has joined #ltsp
12:03twinprism has joined #ltsp
12:03captain_1agnus has joined #ltsp
12:03wwx has joined #ltsp
12:03japerry has joined #ltsp
12:03K_O-Gnom has joined #ltsp
12:03elisboa has joined #ltsp
12:03staffencasa has joined #ltsp
12:03Subhodip has joined #ltsp
12:03indradg has joined #ltsp
12:03mccann has joined #ltsp
12:03mikkel has joined #ltsp
12:03RoninBaka has joined #ltsp
12:03slidesinger has joined #ltsp
12:03jammcq has joined #ltsp
12:03ogra_cmpc has joined #ltsp
12:03popey has joined #ltsp
12:03primeministerp has joined #ltsp
12:03GiantPickle has joined #ltsp
12:03warren has joined #ltsp
12:03jbrett has joined #ltsp
12:03loather has joined #ltsp
12:03rjune has joined #ltsp
12:03savetheWorld has joined #ltsp
12:03ccherrett has joined #ltsp
12:03mistik1 has joined #ltsp
12:03cyberorg has joined #ltsp
12:03Egyptian[Home] has joined #ltsp
12:03sutula has joined #ltsp
12:03Lumiere has joined #ltsp
12:03iMacGyver has joined #ltsp
12:03artista_frustrad has joined #ltsp
12:03xachen has joined #ltsp
12:03jcastro has joined #ltsp
12:03gu has joined #ltsp
12:03gonzaloaf_work has joined #ltsp
12:03|Paradox| has joined #ltsp
12:03klausade has joined #ltsp
12:03loather-work has joined #ltsp
12:03dberkholz has joined #ltsp
12:03ogra has joined #ltsp
12:03Patina has joined #ltsp
12:03topslakr_ has joined #ltsp
12:03sep has joined #ltsp
12:03laga has joined #ltsp
12:03ufo_ has joined #ltsp
12:03vpro has joined #ltsp
12:03leio has joined #ltsp
12:03stgraber has joined #ltsp
12:03moquist has joined #ltsp
12:03daduke has joined #ltsp
12:03alekibango has joined #ltsp
12:03yanu has joined #ltsp
12:03gentgeen__ has joined #ltsp
12:03F-GT has joined #ltsp
12:03MacIver has joined #ltsp
12:03Lns has joined #ltsp
12:03topslakr has joined #ltsp
12:03irule has joined #ltsp
12:03fsgaston has joined #ltsp
12:04warren has quit IRC
12:04primeministerp has quit IRC
12:04elisboa has quit IRC
12:04rjune has quit IRC
12:04GiantPickle has quit IRC
12:04jbrett has quit IRC
12:04ogra_cmpc has quit IRC
12:04loather has quit IRC
12:04japerry has quit IRC
12:04savetheWorld has quit IRC
12:04popey has quit IRC
12:04jammcq has quit IRC
12:04gentgeen__ has quit IRC
12:04F-GT has quit IRC
12:04MacIver has quit IRC
12:04Lns has quit IRC
12:04daduke has quit IRC
12:04alekibango has quit IRC
12:04staffencasa has quit IRC
12:05
<ogra>
there they go
12:05jammcq has joined #ltsp
12:05gentgeen__ has joined #ltsp
12:05F-GT has joined #ltsp
12:05MacIver has joined #ltsp
12:05Lns has joined #ltsp
12:09indradg has quit IRC
12:09staffencasa has joined #ltsp
12:09alekibango has joined #ltsp
12:09daduke has joined #ltsp
12:09japerry has joined #ltsp
12:09elisboa has joined #ltsp
12:09ogra_cmpc has joined #ltsp
12:09savetheWorld has joined #ltsp
12:09rjune has joined #ltsp
12:09loather has joined #ltsp
12:09jbrett has joined #ltsp
12:09warren has joined #ltsp
12:09GiantPickle has joined #ltsp
12:09primeministerp has joined #ltsp
12:09popey has joined #ltsp
12:09Q-FUNK has joined #ltsp
12:10mccann has quit IRC
12:11mhterres has joined #ltsp
12:12Q-FUNK has quit IRC
12:12GiantPickle has quit IRC
12:12rjune has quit IRC
12:12loather has quit IRC
12:12warren has quit IRC
12:12primeministerp has quit IRC
12:12jbrett has quit IRC
12:12popey has quit IRC
12:12elisboa has quit IRC
12:12savetheWorld has quit IRC
12:12ogra_cmpc has quit IRC
12:12japerry has quit IRC
12:13Q-FUNK has joined #ltsp
12:13japerry has joined #ltsp
12:13elisboa has joined #ltsp
12:13ogra_cmpc has joined #ltsp
12:13savetheWorld has joined #ltsp
12:13rjune has joined #ltsp
12:13loather has joined #ltsp
12:13jbrett has joined #ltsp
12:13warren has joined #ltsp
12:13GiantPickle has joined #ltsp
12:13primeministerp has joined #ltsp
12:13popey has joined #ltsp
12:16mhterres has quit IRC
12:17mhterres has joined #ltsp
12:21mistik1_ has joined #ltsp
12:21mistik1 has quit IRC
12:21mistik1_ is now known as mistik1
12:25vagrantc has joined #ltsp
12:26
<ogra>
vagrantc, !
12:26
vagrantc, https://code.launchpad.net/~ogra/ltsp/intrepid-packaging
12:26
(its actually only the hardy packaging copied over yet)
12:29Gadi has joined #ltsp
12:30
<Q-FUNK>
!g
12:30
<ltspbot>
Q-FUNK: "g" is Gadi!!!!!!!!!!!!!!!!!!!!!!!!
12:30
<ogra>
!q
12:30
<ltspbot>
ogra: "q" is Q-FUNK!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
12:30
<ogra>
:)
12:31
<vagrantc>
ogra: thanks! :)
12:31
<ogra>
vagrantc, just fiddling with my flight for next month
12:31
<vagrantc>
ogra: will i see ldm packaging too? :)
12:31
<ogra>
looks like i'll be in boston until 19th
12:32
<vagrantc>
and there's nothing major different in ltspfs, right?
12:32
<ogra>
nope
12:32elisboa has quit IRC
12:32elisboa has joined #ltsp
12:32
<ogra>
and ldm has only added the theme dir and the ldm-ubuntu-themes package in intrepid
12:32
apart from that i took your package
12:32
<vagrantc>
nice.
12:33
<ogra>
(and the themes will go to their own source package later)
12:33mccann has joined #ltsp
12:33
<ogra>
to save bandwith
12:33
we also just had a good discussion how to solve th Xsession problems
12:33
<vagrantc>
oh, cool!
12:33
<Q-FUNK>
:D
12:33
<vagrantc>
major points?
12:34
<ogra>
i'll write up something before i go to bed and send it to -devel
12:34
well, ask ldminfo for Xsession path ... if that doesnt exist use an rc.d script an query the paths
12:34
the merge LDM_SESSION content out of the Xsession path and the actual selected session or put "default" in
12:35
so it works with and without ldminfo support ... without might be a second slower due to checking the paths on the server
12:36
but we can arrange the rc.d script in a way that the first check always matches the build distro ... so connecting to the same distro will be faster, others will be a bit delayed
12:37
<vagrantc>
i was kind of starting to think adding a new LDM_XSESSION thing, so it wouldn't break backwards compatibility ... and a corresponding ldminfo.xsession variable in ldm's c code
12:37
<ogra>
and we can drop all the if statement crap in the C code
12:37
<vagrantc>
yeah, kill that stuff
12:37
<warren>
how would this proposal break backward compat?
12:37
<ogra>
not at all
12:37
<warren>
I really like this proposal
12:37
doesn't need to query if ldminfod gave it what it needed
12:37
<ogra>
if ldminfo doesnt offer the info it will query
12:37
which should work in any case
12:38
<vagrantc>
well, currently LDM_SESSION has one behavior, which replaces the Xsession
12:38
<warren>
and this doesn't break older clients
12:38
<ogra>
yeah
12:38
<vagrantc>
it's original purpose was to actually replace the Xsession, because mdz was using gdm's Xsession script.
12:38
<ogra>
well, you always can set LDM_SESSION anyway
12:38
<vagrantc>
it kind of got grafted on to use it for selecting window manager
12:38* warren tests keyboard layout setting patch
12:38
<ogra>
for clients where it doesnt work
12:38
<warren>
vagrantc: I moved XKB* option handling into ldm itself instead of xorg.conf
12:39
<vagrantc>
but whatever. sometimes it's better to do the right thing when you used to do the wrong thing :)
12:39
<warren>
configure-x.sh should get smaller and smaller
12:39
<ogra>
vagrantc, wrong, mdz called x-session-manager
12:39
everything after that came from me
12:39
<warren>
oh crap!
12:39
ogra: this wont work
12:39
<vagrantc>
ogra: no, i wrote the patch that implemented LDM_SESSION
12:39* Gadi waves
12:39
<ogra>
vagrantc, i mean the switch from x-session-manager to usin Xsession
12:39
<warren>
because if you use something other than ldm it wont set the keyboard options
12:40
<ogra>
the hardcoded part of ldm
12:40
warren, thats why i said initially to put it into the screen.d script
12:40* warren looks at that..
12:40
<ogra>
we can make it a function that can beused by the others as well
12:41
hey Gadiman
12:41
<vagrantc>
yeah, i don't like keyboard handling to be ldm-specific.
12:41
<warren>
ogra: only way to do this is to move x launching from ldm
12:41
ogra: because it has to happen after X launches
12:42* Gadi likes removing X launching from LDM
12:42
<Gadi>
:)
12:42
<warren>
but then if ldm doesn't launch it, then ldm on longer knows the pid of X
12:42
since ldm is expected to restart it
12:42
<Q-FUNK>
:)
12:42
<Gadi>
why should LDM restart it?
12:42
:)
12:42
xinit ldm ....
12:42
when ldm ends, X restarts
12:42
<ogra>
to clear the environment ?
12:42
<Gadi>
easy
12:43
<Q-FUNK>
say, does naybody remember what's the trick to pas kernel cmdline options to the etherboot image?
12:43
<ogra>
yeah
12:43
<warren>
I already wrote the code for ldm to set the keyboard layout from lts.conf
12:43
I could do it again for the other non-ldm screen scripts
12:43
<Gadi>
we dont actually need ldm to launch X
12:43
<ogra>
Gadi, i think there was a reson for that .... we need to ask scott
12:44
<Q-FUNK>
there was one file to set in the chroot's ltsp folder, that is used by update-kernel, but it doesn't seem to affect the etherboot kernel that boot off nbi.img
12:44
<Gadi>
!seen sbalneav
12:44
<ltspbot>
Gadi: sbalneav was last seen in #ltsp 1 week, 1 day, 20 hours, 19 minutes, and 29 seconds ago: <sbalneav> I'll pop in later tonight. Off home for the day.
12:44
<ogra>
which w can do in 4 weeks :)
12:44
<Gadi>
good luck with that ;)
12:44
<ogra>
well, he said he had booked his flight
12:45
<warren>
we need to write down the agenda for the hackfest
12:45
<ogra>
and jammcq stepped on his toes from time to time in the past to make hi show up here
12:45
<Gadi>
just think, if we get rid of launching X, we could run LDM in all sorts of funky X servers
12:45
:)
12:45
<ogra>
warren, localapps
12:45
<warren>
but the wiki at our upstream common site... well... sucks
12:45
<Gadi>
think of the possibilities!
12:45* vagrantc agrees regarding the wiki
12:45* Gadi notes that he is home sick today and prolly a bit dilirius
12:45
<ogra>
Gadi, oooh framebuffer ldm :P
12:46* Gadi was thinking Xnest and friends
12:46* vagrantc proposes to rename it oldwiki.ltsp.org and set up a whole new wiki for wiki.freegeek.org
12:46
<ogra>
warren, at the SF hackfest we actually had a go at the wiki
12:46
<Gadi>
LDM for Windows :)
12:46
hehe
12:46Subhodip has quit IRC
12:46
<warren>
that predates my existence
12:46
<ogra>
we should probably see that people that dont code during the hackfest do some wiki work
12:46* ogra looks at jammcq :)
12:46
<ogra>
warren, nah, that was after detroit
12:47
<Q-FUNK>
where is the hackfest gonna be?
12:47
<Gadi>
Portland, OR
12:47
<ogra>
you just had vanished for a year
12:47
<warren>
We have the server and admin infrastructure to run any wiki engine on its own domain name. (with backups and somebody handling security)
12:47
but it sounds like others can do this as well
12:47
<ogra>
well, we have ltsp.org
12:47
why another wiki ?
12:47
<warren>
the engine at ltsp.org sucks
12:47
<ogra>
lets just see that we get that in shape
12:48
<warren>
and the content is really stale
12:48
<ogra>
oh, right, that swhat you are after
12:48
<Q-FUNK>
portland? again?
12:48
<Gadi>
just keep in mind that many folks still use 4.2
12:48
<ogra>
Q-FUNK, we never had one in portland
12:48
<Gadi>
and that wiki is golden for them
12:48
<Q-FUNK>
any way to hold these in free countries, for a change?
12:48
<ogra>
Gadi, pffft, they wil switch quickly if we steal their docs :P
12:48
<warren>
like china?
12:49
<Gadi>
ogra: nice
12:49
<ogra>
warren, i guess he means estonia :)
12:49
<Gadi>
dont get much freer than Portland (from what I hear)
12:50
<ogra>
apart from being in the wrong country :)
12:50
which you have to enter
12:50
which isnt that easy anymore
12:50
since a month or so
12:51
<warren>
crap, I have to renew my passport
12:51
in a few months
12:51* vagrantc is working on breaking cascadia away from the empire
12:51
<ogra>
with a spoon ?
12:51
*scratch* *scratch*
12:51* vagrantc will not reveal the tools employed
12:52
<ogra>
heh
12:52
<vagrantc>
rebellion can be dangerous work.
12:52
<Q-FUNK>
Gadi: it unfortunately is on the wrong side of the border :(
12:52* ogra still wonders what to do with the 4 spare days between boston and portland
12:52
<warren>
vagrantc: lose the beard first, you'll disappear from their profiling
12:52
<ogra>
i could bring my own spoon and help you
12:52
<warren>
"where did he go!?'
12:53
<ogra>
(if they still allow spoons on planes though)
12:53* Q-FUNK probably looks really weird, IRC'ing on an OLPC in a cocktail lounge in Tallinn
12:53
<ogra_cmpc>
pffft OLPC
12:54
<Gadi>
kids, play nice
12:54
<ogra>
:)
12:54
<Q-FUNK>
yes, uncle Gadi
12:54
<warren>
so...
12:55
ogra: would you be against me just putting setxkbmap into a shell script that is called after launch_x()?
12:55
ogra: and we can discuss how to clean it up at the hackfest
12:55
<ogra>
i like that :)
12:55
no, i wouldnt be opposed
12:55
<warren>
god this sucks, but I'll do it
12:55
so the shell script can be shared with the other screen scripts for now.
12:55* Gadi wonders why setxkbmap cannot be called server side
12:56
<warren>
Gadi: it can, but there's problems with that
12:56
<Q-FUNK>
anyhow, getting back on topic. the kernel in Hardy crashes the DBE60 with a nice kernel oops, right at the start of the nbi.img loading and I need to debug this. one way that the ubuntu kernel team suggested was to boot at a higher resolution, to be able to snap a shot of the full oops.
12:56
<vagrantc>
ogra: whoah. you had to uuencode the .png files in the debian dir?
12:56
<ogra>
Gadi, because japanese people want to type their japanese chars in ldm
12:56
<Gadi>
ah
12:56
<warren>
Gadi: 1) server isn't available before login where you want the keyboard layout set
12:56
<Gadi>
so, its the Japanese now?
12:56
<ogra>
vagrantc, well, patch doesnt take binary data
12:57
<Gadi>
that makes sense
12:57
<warren>
Gadi: 2) setxkbmap is really relayed from the X client to X server and relies on the X server's filesystem
12:57
<vagrantc>
ogra: yeah ...
12:57
<ogra>
and they go into the diff.gz
12:57
thast why i want them in a separate source
12:57
<Gadi>
is this also in an effort to not use xorg.conf?
12:57
<ogra>
but its not urgent ... the themes dont change much anyway
12:57
Gadi, yeah
12:57
<vagrantc>
ogra: i guess... wouldn't it have been easier to separate the source out from the get-go ? :)
12:58
<warren>
Gadi: yes. keyboard layout is the only urgent thing for me without configure-x.sh
12:58
<ogra>
and get rid of the configure-x.sh workaround
12:58
<Gadi>
and this is to what, increase speed?
12:58
boot speed, that is
12:58
<warren>
no
12:58
simplify the config
12:58savetheWorld has quit IRC
12:58
<ogra>
vagrantc, no, it would have taken half an hour or so ...
12:58
<warren>
configure-x.sh is evil
12:58
<Gadi>
simplify what config?
12:59savetheWorld has joined #ltsp
12:59
<ogra>
vagrantc, http://merges.ubuntu.com/main.html still has 52 untouched merges ... merge freeze is tomorrow
12:59
<Q-FUNK>
ogra: which freeze?
12:59
<warren>
ogra: eh... this is simple enough that I'll implement it twice (the ldm version is in C). We can simplify it to the screen.d version later if we agree on how to move X launching away from ldm.
12:59
<ogra>
i can put time in if there is no feeze date approacing me
12:59
<warren>
Gadi: configure-x.sh is so gross that I don't use it.
13:00
<vagrantc>
ogra: does ltsp still depend on libgtk2.0-dev ?
13:00
ogra: er, build-depend ?
13:00
<Gadi>
well, Im sure it thinks you dress funny, too
13:00
<ogra>
Q-FUNK, debian import freeze ... every package on the merge list needs to be touched at least once by then
13:00
<Q-FUNK>
warren: and yet you manage to get a usable keyboard map for every language .. or?
13:00
<Gadi>
but is there a technical reason?
13:00
:)
13:00
<ogra>
vagrantc, oh, my bad i guess
13:00
<Q-FUNK>
ogra: import freeze for intrepid? already?
13:00
<warren>
Q-FUNK: We *are* the cultural imperialists...
13:01
<Q-FUNK>
warren: which is why your culture doesn't reach setomaa :-P
13:01
<warren>
Q-FUNK: well that's the purpose of the screen session itself setting the keyboard layout from lts.conf XKB* options instead of from xorg.conf
13:01
Q-FUNK: (I think the monoculture aspect is especially bad.)
13:01
<Q-FUNK>
why not use the locale for that?
13:02
<ogra>
Q-FUNK, https://wiki.ubuntu.com/IntrepidReleaseSchedule
13:02
<warren>
Q-FUNK: because there is no safe one-to-one mapping of locale to keyboard layout and options.
13:02
<ogra>
Q-FUNK, the locle ?
13:02
nah
13:02
think of the canadians
13:02
<Q-FUNK>
ogra: great. well, I guess that this will be a rare occasion where debian and ubuntu freeze almost at the same time :) hell truly did freeze over :D
13:02* warren began on this keyboard crusade because of one Hebrew user in Israel complaining at me.
13:03
<ogra>
well, and he is right :)
13:03* Gadi knew it would turn to Israel eventually ;)
13:03
<ogra>
but configure-x should have helped
13:03
<warren>
I'm a bit confused by his lts.conf though
13:03
<Q-FUNK>
Gadi: must be the zionist conspiration again :-P
13:03
<ogra>
Gadi, i brought up canada deliberately :)
13:03
<savetheWorld>
all we have to do is officially eliminate any language with less than 100 million native speakers....
13:04
<warren>
XKBLAYOUT = "us,il"
13:04
you can have two?
13:04
<savetheWorld>
that'll reduce the workload significantly.
13:04
<warren>
(this doesn't make sense)
13:04
<ogra>
warren, yes, but you need as switch option as well
13:04
s/as/a/
13:04
<warren>
he also has
13:04
XKBVARIANT = ",si1452"
13:04
setxkbmap doesn't like this
13:04
<ogra>
that doesnt look right
13:05
<warren>
I wonder if this was tripping some bug in X's config parser that made it work.
13:05
=)
13:05
<Q-FUNK>
savetheWorld: you do have a sense of irony, for someone with your nickname
13:05
<ogra>
i dont know the exact pattern from the top of my head but the switch needs to be in XKBOPTIONS iirc and looks quite different
13:05
<savetheWorld>
yes, yes i do.
13:05
warren: or tripping some undocumented feature?
13:06
<warren>
the entire internals of X are undocumented pretty much
13:06
<ogra>
well, you have the code :)
13:06
<savetheWorld>
eep!
13:06
<ogra>
who needs docs with good code
13:07
<Gadi>
who needs good code when you have docs?
13:07
:)
13:07
<vagrantc>
who needs bad docs when you have obfuscated code?
13:07
<ogra>
who needs code at all
13:07
<vagrantc>
hear hear!
13:08
i think we've found the problem.
13:08
i feel kind of dense for not picking up on it before.
13:08
<ogra>
heh
13:08
<Q-FUNK>
good code tends to be human-readable. it also tends to be non-existent on this planet.
13:09
I really wished that the OpenBSD guys would enter the X.org universe...
13:10
"not a single root window exploit in 10 years"
13:11indradg has joined #ltsp
13:16
<vagrantc>
on a server that's got rather limited uses (i.e. a default OpenBSD install)
13:20
<Q-FUNK>
it's jsut that we need the sort of rock-solid refactoring that OpenBSD are good at, to really make X.org usable and reduce its memory consumption.
13:20
<ogra>
heh, you should come to #ubuntu-server and tel that to every second person entering there asking how to install gnome :)
13:25
<warren>
Ubuntu server doesn't have GNOME?
13:26
<ogra>
nah
13:26
<Q-FUNK>
ogra: according to http://packages.ubuntu.com/search?keywords=xserver-xorg-video-geode intrepid still doesn't have the latest imported from debian
13:26
<ogra>
who needs a server with gui ?
13:26
<Q-FUNK>
ditto for http://packages.ubuntu.com/search?keywords=xserver-xorg-video-nsc
13:27
<ogra>
as long as it was synced once all is fine
13:27
deban import freeze means only that all manual merges need to be done
13:27
<Q-FUNK>
those currently have ubuntu ddiffs. it might have prevented the sync
13:27
<ogra>
and dont rely on p.u.c ...
13:28
its usually behind
13:28
<Q-FUNK>
well, there's no ohter reliable source of info...
13:28
<ogra>
sure
13:28
LP or archive.ubuntu.com
13:29
there is no other *convenient* source :)
13:29
<Q-FUNK>
argh
13:29
<ogra>
LP carries whats supposed to be in the archive ... a.u.c is the only actually reliable source and p.u.c is the source that tells you what was in yesterday or so
13:30
but as i said, doesnt matter for the freeze for your packages
13:30
<Q-FUNK>
ok. LP shows that intrepid is up-to-date. however, hardy still hasn't been resolved. it's really starting to annoy me.
13:30
<ogra>
they were both already uploaded to intrepid
13:30
<Q-FUNK>
yes, intrepid is verified as OK according to LP
13:31
<ogra>
it just for us to see that we have all manual merges done
13:31
and its surely not as strict as other freezes ... (i wouldnt know whom to ask for a merge freeze exception even)
13:31
but its a deadline nontheless
13:32
and hotkey-setup is "§%E§&$ !!!
13:32
grmbl
13:32
<Q-FUNK>
Hardy is what worries me, as it's supposed to be LTS. still, with all the quirks in it, lots of people like Gadi decided to skip Hardy, which is not s good indication of ubuntu's overall quality.
13:33
<ogra>
well, users where told to wait for 8.04.1 in corporate environments for a reason :) thast why we had that in the release notes
13:34
<Q-FUNK>
right, but in ubuntu, whatever goes to hardy-updates doesn't end up pushed in hardy. they remain separate repositories, which requires addtional backflips to generate the LTSP chroot
13:35
in debian, whatever was uploaded to security over the last few months gets merged into main at the next point release.
13:36
<ogra>
it ends up on the CD
13:36
and hardy-updates is on by default on every hardy install
13:36
<Q-FUNK>
plus, filing an SRU in time doesn't guarantee that it will be approved on time for 8.04.1, which really sucks.
13:37
<ogra>
well, depends how your SRU looks
13:37
<Q-FUNK>
if nobody gets around approving it, it doesn't matter how long it has been open, it misses the deadline
13:37* ogra had about 6 different SRUs that directy went through
13:37
<Q-FUNK>
not really. like my initial fix for geode was open before the 8.04.0 freeze, but it wasn't checked until after the freeze
13:38
<ogra>
yeah, but you still have no simple patch
13:38
<Q-FUNK>
it's n ot called an SRU for nothing: it required a new upstream, not a simple packaging patch.
13:38
<ogra>
but it should have been the fix only
13:39
while you pushed for a full package upgarde
13:39
<Q-FUNK>
which required a new upstream.
13:39
<ogra>
so that indeed delays everything
13:39
<Q-FUNK>
yes, because fixing hardware freezes required a new upstream, plus updated packaging
13:40
at least, pitti finally accepted my explanation as to why we must diverge from the file structure used in 8.04.0
13:41
what remains to be seen is whether slangasek will agree on time to allow a freeze exception, now that the while geode+nsc issue remained unchecked for so long
13:41
<ogra>
slangasek will follow up, i'm sure
13:41
but you had several fixes you pushed through that didnt fix it and one that even had to be reverted, so indeed they are cautious
13:42
<Q-FUNK>
I was glad to finally meet him in person, at UDS. we even ended up eating together, before the club party. it has put him in a whole other perspective.
13:42
<vagrantc>
we need to implement pid files for cdpinger...
13:42
<ogra>
(due to the pci id crap)
13:42
vagrantc, why is that ?
13:43
<vagrantc>
ogra: well, i want it to be able to check if cdpinger is already running with the same argument
13:43
<Q-FUNK>
yes and no. the main reason why the fix that is currently in prposed had to be fixed again is becuase pitti initially insisted that I should fit everything within the file structure used in 8.04.0
13:43
<ogra>
vagrantc, ah ... cant we call it with daemon tools ?
13:43
<Q-FUNK>
my initial fix for this using 2.9.0 from debian, backported to hardy, worked out of the box.
13:43
<vagrantc>
i suppose so.
13:44
my other idea was to simply write a wrapper script, and move cdpinger into libexec
13:44
the real cdpinger
13:44
<ogra>
Q-FUNK, right but doesnt comply with the SRU rules
13:45
<vagrantc>
i'm trying to do the cdpinger from udev thing again ... i didn't like how my last attempt turned out.
13:45
<Q-FUNK>
ogra: cherry-picking is more error-prone than simply building an existing solution using older build-depends.
13:45
<ogra>
well, but not complying to the rules and then complaining that we suck simply isnt right
13:45
<Q-FUNK>
anyhow. learning how to work within an SRU was a nice learning experience, though one I'd rather not repeat.
13:45
<ogra>
yeah, i belive you :)
13:46
<warren>
btw, how does the rdesktop screen script work?
13:46
it doesn't appear that anything launches x
13:46* ogra never tested it
13:46
<Q-FUNK>
well, the solution was more complex than I myself wlould have bargained for. I would have preferred to just do a small patch that fixes the packaging, but I also had to package a new usptream to fix the real hardware issue.
13:46
<ogra>
but i know theer are people using it successfully
13:47
<Q-FUNK>
...plus i had to learn about debian's nasty, non-upstream compliant way of doing PCI ID matching
13:47
<ogra>
Q-FUNK, indeed i understand and i doubt i would have solved it better or more elegantly but still, there is a process that wasnt followed so it took longer
13:48
i'm pretty sure the fix will be in 8.04.1
13:48
<Q-FUNK>
the only comforting aspect is that pitti was extremely supportive, all while being very strict about folowing proper procedure :)
13:48
<ogra>
steve wont block
13:48
<Q-FUNK>
hopefully :)
13:49
I think that he has learned to appreciate me a bit more at UDS
13:49
<ogra>
did you know each other from debian ?
13:49
<Q-FUNK>
before that, on the debian side, he has been more abrasive than necessary towards me, too many times
13:50
yeah
13:50
he's essentially the guy that not only crashed my NM procees but also got me banned from being added to DM keyring
13:51
he has a way of being confrontational when he really means to be merely inquisitive and thorough.
13:51
it has gotten us on the wrong foot too many times.
13:51
<warren>
how long until 8.04.1?
13:52
<Q-FUNK>
warren: we're currently in freeze already, afaik
13:52
<vagrantc>
clearly, it will be released 2008, april 1st.
13:52
<warren>
will there be new install media for it?
13:52
<Q-FUNK>
IIRC 8.04.1 is due out sometimes in July
13:53
AFAIK slangasek and mpitt are currently hand-checking a few last-minute freeze exceptions this week and then it's a release.
13:57
<ogra>
warren, proposed date is july 3
13:57
Q-FUNK, doesnt look like slangasek is happy with nsc
13:57
<warren>
interesting...
13:57
setxkbmap actually causes X to die
13:58
<Q-FUNK>
ogra: why not?
13:58
<ogra>
look at the bug :/
13:58
<Q-FUNK>
ogra: is he a nnywhere online where I can discuss this and find a compromise?
13:58
<warren>
nsc X driver?
13:58
<ogra>
#ubuntu-devel
13:59gdi2k has quit IRC
13:59
<ogra>
warren, well, a backport of that to the stable release
13:59
warren, nsc in itself is fine
14:00
but we have a policy that only the most minimal change is allowed in a stable release and it seems Q-FUNK backported debian packaging as well
14:00
<warren>
oh, I don't want to use ldm_spawn for setxkbmap...
14:11
<Q-FUNK>
ogra: he replied only a few minutes ago. I just added my bit now.
14:12
<ogra>
well, he'S in #ubuntu-devel
14:12
in case you want to discuss more directly
14:12
<Q-FUNK>
warren: debian has a non-upstream wahy of doing PCI ID matching that's been verhy painful tlo work with, because some of the PCI ID are shared by cyrix, nsc and geode drivers.
14:13
<warren>
ah yes
14:13
you need more fine grained than that
14:13Pascal_1 has joined #ltsp
14:13
<warren>
merging nsc into geode is going to be the only way to clean this up fully
14:13
<ogra>
we had more fine grained than that in ubuntu
14:13
<warren>
If I had nsc hardware I'd try to help
14:13
<Q-FUNK>
debian _has_ more fine-grained, but at the same time, it's too random
14:14
<ogra>
and then our X maintainer resigned and we took the debina packages plain as they are
14:14
<warren>
who was the X maintainer?
14:14
<ogra>
and now all ubuntu packaging is based of the debian one
14:14
daniel stone ...
14:14
he's at nokia now
14:14Pascal_1 has quit IRC
14:14
<ogra>
posting a lot on the hal list
14:14
<Q-FUNK>
the main thing is that debian auto-generates PCI ID list by parsing the source code for any occurence of PCI_VENDOR and PCI_DEVICE. it ends up generating non-existent numbers, and duplicates.
14:15
<warren>
wow
14:15
that's bad
14:15
<Q-FUNK>
to make matters worse, the source code was esentially forked every time the geode product line changed owner, but kept its ancestral roots
14:16
so, all 3 drivers still have traces of cyrix heritage
14:16
plus nsc and geode both have the nsc heritage too
14:16
<warren>
I wonder, how do we obtain NSC geode hardware?
14:17* warren looks on ebay
14:17
<Q-FUNK>
DBE60 :)
14:17
debian replied to my bug report by saying that it's not their fault if all 3 drivers "claim" support for the same hardware.
14:18
<warren>
having unused code is not a claim...
14:18
<Q-FUNK>
this is bullshit, since the code itself has conditional matching for specific vendor+device combinations, but the debian awk rule blindly matches ANY damn PCI vendor and device ID found in the source.
14:19
e.g cyrix and nsc have a device ID for a bridge chip that is needed to access the video hardware, so the debian point and shoot script claimed support for a "video" chip that has an IDE controller!
14:20
precisely
14:20
and yet they call unused code "hardware support claim"
14:20
anyhow...
14:20
<warren>
I think our default behavior with that hardware is only slightly better
14:20
"I don't recognize this thus I wont work"
14:21
so is Ubuntu going to fork it or something?
14:21
<Q-FUNK>
all this made debuging and fixing not just the -geode issue but, as it turns out, the -nsc issue too, a LOT more difficult than I had bargained for and the resulting fix a lot more involved than what SRU would normally allow.
14:22
<warren>
SRU?
14:22
<ogra>
stable release update
14:22
<Q-FUNK>
which is what created this mess, to fix hardy, which is supposed to be a stable long-term release.
14:23
<ogra>
warren, i'll bring you a rub o tattoo with the abbreviation to portland :) its the third time you ask ;)
14:23
s/o/on/
14:23
<Q-FUNK>
in principle, the SRU policy makes perfect sense; you indeed want to avoid intrusive changes, for a stable release, to reduce the risk of a regression creeping in.
14:23
:D
14:23
<warren>
stable == unchanging
14:24
<ogra>
Q-FUNK, well, you patch comments and stuff ...
14:24
or patch in lies that are then commented
14:24
<Gadi>
"Taming of the SRU"
14:24* Gadi couldnt resist
14:24
<ogra>
i can understand that steve doesnt see that as a minimal patch
14:24
<Q-FUNK>
in practice, trying to fix 3 packages, plus a hardware issue reported by a user which required packaging a new geode upstream, turned into this horrible monster that was difficult to fix to everyone's satisfaction.
14:25
ogra: I mostly agree, but this was a complex issue to fix. it would have been more easily accepted if handled by someone else than me.
14:26
<warren>
it sounds like the core algorithm from debian is broken
14:27
<ogra>
warren, well, you guys call redhat stable as well and even introduce new kernels :)
14:27
or did i miss that you call RHEL "moving target" :)
14:28
<warren>
I don't claim anything
14:28
I'm not in that part of the company
14:28
<ogra>
oh ?
14:28
<warren>
I don't work on RHEL?
14:28
<ogra>
but for redhat :)
14:28
<warren>
we have people working on different things that have nothing to do with RHEL
14:29
or little to do
14:29
<ogra>
like canonical
14:29
80% of us dont work on the distro
14:30
until a year ago the distro team was actually 18ppl ...
14:31
i think we roughly doubled up now ... but still your distro team is bigger than my company
14:43slidesinger has quit IRC
14:57mikkel has quit IRC
14:58elisboa has quit IRC
15:18Q-FUNK has quit IRC
15:19
<ogra>
phew, so seems Q-Funk might still get his SRU through if he manages to fix the packages tonight
15:24
<warren>
I have all the C code to do this, but I'm having trouble getting the environment variables to the forked process.
15:25
or setxkbmap is looking for something else that isn't getting passed ...
15:25
<ogra>
i tink g_spawn_async inhertis the env
15:26
<warren>
yeah, I guess I should use glib instead of raw fork
15:26
<ogra>
yeah
15:26
we ave it as dep anyway and is adds some saftey net
15:26
*have
15:26
<warren>
shouldnt' I use sync instead of async?
15:27
<ogra>
i think that blocks the main process
15:27
which is not what we want
15:27
<warren>
one issue here...
15:27
X was started async
15:27
we're setting keyboard immediately after
15:28
X might not be sufficiently started up to do the keyboard
15:28
right?
15:28
<ogra>
oh, right
15:28
add a sleep or so
15:28
<warren>
fork and sleep might be our only option
15:28
<ogra>
(the the externa script, not the C code)
15:28
*external
15:29
<warren>
sleeping is no guarantee though
15:29
<ogra>
indeed, only a workaround
15:30
<warren>
hmm
15:30
<ogra>
and a matter of luck as well
15:30
<warren>
only safe way to do it, is to set keyboard layout AFTER an X client has started
15:30
like the greeter.
15:30* ogra just remembers the sound issue with very fast clients
15:30
<ogra>
do it *from* the greeter probably ?
15:30
<warren>
um..
15:30
that might be a goodi dea
15:30
good idea
15:30
<ogra>
nobody will type anyway before thats up
15:31
<warren>
greeter doesn't have access to ldmlog though
15:31
<ogra>
does it need to ?
15:32
apart from deugging
15:32
*debugging
15:32
<warren>
right
15:33
<ogra>
but all you need is the spawn and call a script from that
15:34
the rest is scripting
15:34ace_suares has quit IRC
15:35Pascal_1 has joined #ltsp
15:36indradg_ has joined #ltsp
15:40Nuba1 has joined #ltsp
15:41Nubae has quit IRC
15:41indradg has quit IRC
15:42
<warren>
ogra: hm, spawning a shell script doesn't work with setxkbmap either
15:42
I tried exporting both DISPLAY and XAUTHORITY and that doesn't seem to be enough
15:42
setxkbmap tries to connect to the X server
15:43
<ogra>
hmm
15:43
<vagrantc>
it must be run client-side ?
15:43
ah yes, to get ldm working
15:43
<ogra>
right
15:43
before the PW input field is there
15:43
or *when* its there
15:43
<vagrantc>
and username :)
15:44
<ogra>
indeed :)
15:44
this input thing
15:45
<warren>
god I'm really hating keyboard now
15:45ace_suares has joined #ltsp
15:45
<warren>
This might drive me to support the imperialist monoculture.
15:46|Ryan52 has quit IRC
15:46
<vagrantc>
i'm almost tempted to add a recommends ltsp-client-core to ltspfsd ... just to be able to properly handle /usr/share/ltsp/ltsp_config and /usr/share/ltsp/ltsp-common-functions
15:46
<ogra>
and use windows ?
15:47
vagrantc, didnt we want that anyway at some point ?
15:47
for tcos
15:47
<vagrantc>
ogra: oh, you mean the package split ... last i recall it was everybody but me who wanted that :)
15:48
<ogra>
ah, k
15:48
<warren>
I think I have to dig into setxkbmap and X source in order to understand what it wants.
15:48
<vagrantc>
but i'm starting to come around
15:48
<ogra>
to be honest i dont really care but it was a way to fix his issue back then
15:49
he didnt really follow up on it either
15:51
oh my ... germany is in the finals of the euro soccer championship
15:51
<laga>
yay
15:51* ogra hears hell break loose outside
15:51
<laga>
yay
15:51
<warren>
last week I saw everyone go nuts here
15:51
they were wearing green jerseys
15:51
and were drunk at noon
15:52
I had no idea what was going on.
15:52
<ogra>
green ?
15:52
i thought boston is red socks territory
15:52
<warren>
apparently some basketball
15:52
<ogra>
ah
15:52
<warren>
Boston Celtics
15:52
<ogra>
there are other sports as well :)
15:52
<jammcq>
heh, "some basketball" ???
15:52
<warren>
I don't follow any sports
15:53
<jammcq>
the celtics won their 18th NBA championship
15:53
<warren>
I don't care about the Red Sox, Celtics or Patriots
15:53* ogra neither
15:53
<ogra>
but getting around soccer in germany isnt easy on days like today
15:53
<jammcq>
basketball is one of the few sports I do follow
15:54|Ryan52 has joined #ltsp
15:54
<warren>
it took me 1.5 hours to take the subway what normally takes 30 minutes
15:54
<ogra>
i'm waiting since 1.5h for a pizza :)
15:54
GF assumes it will still come ... have given up
15:54
s//i/
16:06* warren looks at setxkbmap source
16:07mhterres has quit IRC
16:08
<warren>
hmm, setxkbmap only wants DISPLAY
16:08
<vagrantc>
anyone see a problem starting ltspfsd from ltspfs_entry hooks in udev rules ?
16:08
<ogra>
not at all
16:09
<vagrantc>
i've proposed all this stuff ages ago, but suspect it hasn't really gotten much review :)
16:11
<ogra>
i never objected
16:11
<vagrantc>
:)
16:11
<ogra>
even when we were discussing about add/remove_fstab_enrty
16:12
:)
16:12
well, if oyu add it upstream *now* it will get testing in ubuntu and you can safely pull it into debian if thats successfull
16:13
i'm allowed to break the world in intrepid atm and test out stuff ... so feel free to throuw stuff at me
16:14
<vagrantc>
well, i've tested this stuff moderately well myself, so i'll just go ahead and push to upstream and debian and you can pull it if you want.
16:14
<ogra>
debian and me ?
16:14
<vagrantc>
well, more likely ubuntu and you :)
16:14K_O-Gnom has quit IRC
16:15
<ogra>
:)
16:15
well, i'm pondering to take over xaos
16:15
in debian
16:15
<vagrantc>
but *you* could pull *from* debian *into* ubuntu :)
16:15
<ogra>
i maintain it in ubuntu anyway and joey orphaned it after 15 years
16:16
so makes a good "foot in the door" package
16:18
well, you are to freezed atm i guess to pull in the freaky testing stuff into lenny
16:18
so i'll better go with upstream atm
16:19
<vagrantc>
you talking xaos or something else?
16:19
<ogra>
the last two lines were ltsp :)
16:19
<vagrantc>
we're not really frozen, that's just the udeb stuff
16:19
<ogra>
ah
16:20
<vagrantc>
speaking of which, i should try and merge your intrepid branch...
16:21
<ogra>
that needs cleanup
16:21
<vagrantc>
i worked a bunch of udeb the other day, but still haven't gotten it in working condition
16:21Pascal_1 has quit IRC
16:21
<ogra>
i just pulled over the hardy stuff and took a fresh upstream tarball
16:21
<vagrantc>
no changes since the package in hardy? i just grabbed the script from the sources published in hardy...
16:21
<ogra>
i dont even know if it builds or works
16:21
s/builds/builds a client/
16:22
no packaging changes, no ... some minor stuff like moving lts-parameters.txt into server
16:22
but nothing really notable
16:23
<vagrantc>
ogra: you pushed a packaging branch, though, right?
16:23
<ogra>
the one i gave you when you joined
16:24
<vagrantc>
yeah, cool.
16:24* ogra starts getting desperate with the hotkey-setup merge
16:24
<ogra>
kudos to pere for maintaining it
16:25
<vagrantc>
desperate merges, that sounds fun.
16:25
<ogra>
well, i dont really have a clue about the package
16:25
i can work out the merge conflicts and all, but dont know if it works right
16:26
<vagrantc>
ogra: ah, so you just created a fresh packaging branch ...
16:26
<ogra>
matthew garret going to redhat was a major loss for ubuntu and debin
16:27
well, i bzr'ed the debian dir
16:27
after copying it over from hardy
16:27
<vagrantc>
sure.
16:27
so it has no merge history...
16:27
<ogra>
ah, no
16:28
<vagrantc>
i actually have the full history ... right back to revision 1 :)
16:28
<ogra>
lol
16:28
and even that has a typo in the commit
16:29Q-FUNK has joined #ltsp
16:29
<vagrantc>
so, basically, i can't really merge this branch except manually
16:29
<ogra>
because there is no common ancestor ?
16:30
<vagrantc>
yeah
16:30
or is there some way to trick it?
16:30
<ogra>
hmm, i dont think so
16:31
<vagrantc>
it does talk about base revisions ...
16:39Egyptian[Home] has quit IRC
16:41Egyptian[Home] has joined #ltsp
16:42irule has quit IRC
16:44
<ogra>
Q-FUNK, so bryce just uploaded -nsc
16:44
we need testing and bug feedback
16:45
<Q-FUNK>
ok
16:45
it's in proposed?
16:45
<ogra>
i triggered cr3
16:45
well, just uploaded ....
16:46
<Q-FUNK>
is the updated geode also there?
16:46
<ogra>
it will be there in 1-2h
16:46
i think its only geode yet
16:46
err
16:46
nsc
16:46
didnt ou say you wanted t fix the changelog there ?
16:46
<Q-FUNK>
we need both, because of the ID conflict
16:47
or actually, all 3 packages
16:47
what about cr3? for testing against some hardware?
16:47
<ogra>
yeah, he has the koolu stuff in the testlab in montreal
16:47
and i guess some other geode systems
16:48
<Q-FUNK>
he has a thincan and the koolu
16:48
<ogra>
ah
16:48
i didnt know about the thincan
16:48
<Q-FUNK>
this is good for testing half the geode but not much else.
16:48
<ogra>
he never told me
16:48
<Q-FUNK>
for testing nsc, you have a dbe60 :)
16:49
<ogra>
i know what i have ... i didnt know he had
16:49
<Q-FUNK>
I gave him one to bring back at UDS, so that he could A/B test with the koolu, on any eventual geode issue
16:49
however, few people have any nsc GX1 or nsc/amd gx2, except for Gadi
16:50
you're a rare exception to have any GX1 hardware and thta only covers the NSC era
16:50
<ogra>
Gadi, https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/219630 help, we need testers
16:50
(and comments on the bug)
16:50
<Q-FUNK>
for GX2, Gadi is the only one I know who has any
16:51
<warren>
hmm
16:51
I have a GX something dev board here
16:51
but no case for it
16:51
<ogra>
and no hardy install
16:51
i bet
16:51
<warren>
yeah, but I need the hardware setup if I want to help in merging the driver.
16:52
<ogra>
well https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/219630 in case you test ubuntu stuff on it :)
16:52
<warren>
well, given that I have no case, PSU, etc.
16:52
gotta transit
16:53
<Q-FUNK>
for LX, that's 80% of the market and everyone has hardware from at least two vendors on hand
16:53
GX1 hardware dating back from the Cyrix era is the big question mark. I don't anyone who has any
16:53
oh, wait. dilinger has access to hardware from all era, at the OLPC labs
16:54
<Gadi>
does anyone need GX2 hardware for testing?
16:55
I am happy to get someone a GX2 system if it leads to driver improvements :)
16:55
<vagrantc>
so, my cdpinger from udev stuff basically works ... but i either 1) move /usr/sbin/cdpinger to /usr/$libexec/ltspfsd/cdpinger and make a wrapper script in /usr/sbin/cdpinger that figures out the appropriate symlink name or if it's already running .... 2) implement all that in cdpinger itself
16:55
<ogra>
Gadi, well, we only need a single test of the packages in hardy-proposed so they can get on the 8.04.1 CD
16:56
<vagrantc>
|Ryan52: rearing to do some C hacking? :)
16:56
<ogra>
i dont want your HW here lying around
16:56
<|Ryan52>
vagrantc: maybe.
16:56
<Q-FUNK>
ogra: I would mainly need to know where to download the updated packages from, so that I can test them tomorrow morning
16:57
<vagrantc>
i have the cdpinger wrapper script... then i need to wrestle with automake to install cdpinger to /usr/$libexec/ltspfsd/ ....
16:57
<ogra>
Q-FUNK, well, hardy-proposed
16:59
Q-FUNK, oh, and it seems the nsc package you uploaded FTBFSed, seems you had a patch in debian/patches that was already in the upstream source
16:59
so it broke on that
16:59
(bryce commented on the bug)
17:02
<Q-FUNK>
strange, because it built fine on the PPA
17:02
yes, I just noticed now
17:02
<vagrantc>
|Ryan52: well, the bzr branch http://bazaar.launchpad.net/~ltsp-upstream/ltspfs/ltspfs-trunk/
17:02
|Ryan52: src/cdpinger.c ...
17:03
|Ryan52: needs to know if an existing cdpinger is running with the same arguments ...
17:04
|Ryan52: and needs to figure out which symlink to use ... i.e. cdpinger /dev/hdc ... should re-exec itself as cdpinger /dev/cdrom or /dev/cdrom0 depending on which symlink it was pointng to ...
17:05
<|Ryan52>
okay. (you mean cdpinger /dev/cdrom should re-exec itself as cdpinger /dev/hdc, right?)
17:06
<vagrantc>
|Ryan52: other way around ...
17:06
<Q-FUNK>
well, it seems that he tried to merge in the debian changes without removing the old patch first
17:06
no wonder it FTBFS
17:06
his debdiff makes sense
17:06
I'll fetch the source and check
17:06
<vagrantc>
|Ryan52: there doesn't appear to be a simple way to figure out corresponding symlinks from udev
17:07
<|Ryan52>
vagrantc: huh? I don't get it...why would you do that?
17:09
<vagrantc>
|Ryan52: the information we can get easily from udev is there is a device named /dev/hdc that is a CD device ... but we'd rather use the symlink'ed name because "hdc" doesn't make for a good CDROM icon name.
17:09
<|Ryan52>
ah. okay.
17:09
<ogra>
vagrantc, uh, why so complicated ?
17:09
<vagrantc>
maybe i should just handle the symlink stuff in the udev hook.
17:09
<ogra>
lets just rely on udev and use /dev/cdrom
17:10
<vagrantc>
ogra: because udev never tells us anything about /dev/cdrom
17:10
<ogra>
but it creates the link proper
17:10
<vagrantc>
ogra: we have to call udevinfo (or udevadm info) to get that
17:10
ogra: also to properly handle cdrom0, cdrom1, etc.
17:10
<ogra>
pfft
17:11elisboa has joined #ltsp
17:11* ogra is totally ignorant wrt multiple cdroms on thin clients
17:11
<vagrantc>
|Ryan52: actually, if you can just get cdpinger to recognize if it's already running under the same name, i think the symlink stuff is easier handled elsewhere.
17:11
<|Ryan52>
okay.
17:12
<ogra>
vagrantc, if you urgently want cdromX lets use an override udev rule, stolen from the original cdrom handling rule and just added the cdpinger command
17:13
<vagrantc>
ogra: i'm open to clean suggestions ...
17:14
ogra: but so far, the only way i've figured out is to call udevinfo -q symlink -n $device
17:14
<ogra>
i'm in a meeting alongside this discussion, but i think we can easily have a udev rule to handle it
17:18
<vagrantc>
i guess we could use PROGRAM and pass $result or %c
17:23juanpaul has quit IRC
17:24juanpaul has joined #ltsp
17:24johnny2 has joined #ltsp
17:26
<vagrantc>
ogra: i remember why i didn't want to do it directly from udev ... makes it harder to respect LOCALDEV
17:26
<ogra>
isnt LOCLDEV globally exported ?
17:27
i thought it was
17:27
<vagrantc>
to udev ?
17:27
<ogra>
well, no that runs to early
17:27
<vagrantc>
globally exported to all scripts that source ltsp_config ...
17:27
<ogra>
but to a script it clls
17:28
<vagrantc>
by putting it in ltspfs_entry ... we can easily add all the LOCALDEV checking in one place.
17:30
<ogra>
yeah, so lets call ltspfs_entry for cdroms then from a udev rule
17:35indradg_ has quit IRC
17:37leio has quit IRC
17:37leio has joined #ltsp
17:40
<vagrantc>
right
17:41
that's what i'm working on now ...
17:41
<ogra>
cool
17:41
<vagrantc>
i've tried this half a dozen times ... and it always looks uglier than i initially think it'll be.
17:41jammcq has quit IRC
17:42
<vagrantc>
like, i get a revalation how to do it simpler, and it never pans out to be simpler
17:47japerry has quit IRC
17:47supreme_ has joined #ltsp
17:47
<supreme_>
hi all
17:47
how can i fix maximum concurrent sessions reahced problem?
17:47
i have like 20 clients
17:48Q-FUNK has quit IRC
17:50
<supreme_>
i think i fixed thanks
17:50
<ogra>
where do you see that ? we dont set any limit, ldm will just use ssh so whatever is set up in ssh will do
17:50Q-FUNK has joined #ltsp
17:50johnny2 has left #ltsp
17:50johnny2 has joined #ltsp
17:51japerry has joined #ltsp
17:52
<supreme_>
ogra, this problem happen when i power on more tha like 20 clients
17:52
but in /etc/X11/gdm/gdm.conf i can fix this
17:52
ill try
17:52
<ogra>
gdm ?
17:53
oh, you were the one with the ebox who used 4.2 ...
17:53
<|Ryan52>
vagrantc: What am I supposed to do to get a Makefile?
17:53* ogra remembers now
17:53
<vagrantc>
|Ryan52: ./autogen.sh ... ./configure ?
17:54
<|Ryan52>
from autogen.sh:
17:54
Makefile.am: required file `./ChangeLog' not found
17:54supreme_ has quit IRC
17:56
<vagrantc>
touch ChangeLog ... i guess.
17:57
<|Ryan52>
oh, okay...
18:02
<vagrantc>
i've got an acceptible workaround for the moment, i think.
18:07
but i'm curious if it can be more cleanly/maintainably handled in C
18:09japerry has quit IRC
18:09japerry has joined #ltsp
18:17japerry has quit IRC
18:17captain_1agnus has quit IRC
18:17captain_magnus has joined #ltsp
18:20
<vagrantc>
committed, but it's ain't the pretties code ...
18:20
it's ain't. i love my english language.
18:35
<Gadi>
btw, vagrantc: I have also added some lts.conf params for ltspfs mount options and device type excludes and such
18:36
remind me to merge some of that in once you finish your rework
18:36
I find it very useful to have ltsp_config called in the ltspfs_entry stuff
18:37
<vagrantc>
Gadi: well, one of my most recent changes sources ltsp_config :)
18:38
<Gadi>
yeah - I do that already
18:38
doesnt add time at all
18:38
<vagrantc>
well, maybe not the way *you* do it :)
18:38
<Gadi>
what do you mean?
18:39
<vagrantc>
i feel like my last three commits were perhaps more complicated than they needed to be, but probably not overly evil performance-wise.
18:39
<Gadi>
heh
18:39
well, we can streamline stuff together in Portland, if you like
18:39
I hacked up ltspfs_entry so much
18:39
its ugly
18:39
<ogra>
tsk
18:39
<Gadi>
:)
18:40
but, then, Ive hacked up a bunch of things
18:40
mostly to support my -er -local ltspfs
18:40
:)
18:40* ogra twiddles thimbs waiting for LP
18:40
<ogra>
thumbs as well ...
18:40
<Gadi>
but, some useful things - like mount options let you say "mount everything read-only"
18:41
even if its a read-write drive
18:41
(useful for security in some environments)
18:42
<vagrantc>
Gadi: ooooh.
18:42
<ogra>
do you have any hack to use local haddrives ?
18:42
<Gadi>
i know not very earth-shattering
18:42
<ogra>
that comes up often recently
18:42
<Gadi>
but useful to some
18:42
<vagrantc>
i like the idea of adding mount options ... like noatime and such for filesystems that support it.
18:42
<Gadi>
ogra: that usually amounts to ntfs
18:43
<ogra>
Gadi, well, no probelm in ubuntu
18:43
<Gadi>
well, except the driver is not loaded by default
18:43
<ogra>
vagrantc, doesnt debians kernel default to relatime anyway ?
18:43
<Gadi>
(in the chroot)
18:43
<ogra>
Gadi, its loaded by initrmafs by default
18:43
*initrmafs
18:43
bah
18:43
<Gadi>
since which distro?
18:44
not in gutsy
18:44* ogra gives up correcting himself
18:44
<Gadi>
afaik
18:44
<ogra>
sure in gutsy
18:44
thats why we have fuse loaded by default fro initramfs
18:44
<Gadi>
ntfs?
18:44
<ogra>
ntfs3g
18:44
<Gadi>
hmm
18:44
<ogra>
or however that thing is called
18:44
<Gadi>
are you positive?
18:45
<ogra>
yes
18:45
<Gadi>
I feel like the answer to everyone in this channel looking for it is: echo "ntfs" >> /opt/ltsp/i386/etc/modules
18:45
:)
18:45
<ogra>
i added parts of it
18:45
eeek
18:45
dont do that
18:45
<Gadi>
and that usually works
18:45
lol
18:45
<ogra>
and might do evil things to your data
18:46
<Gadi>
so, what prob do folks have with local hard drives?
18:46
<ogra>
really dont use the ntfs module
18:46
<Gadi>
(well, its ro by default)
18:46
<Q-FUNK>
ok, -all just entered hardy-proposed and will probably be pushed as-is into hardy-updates
18:46
<ogra>
nsc was updated as well
18:46
<Q-FUNK>
Depends amd -> geode
18:46
<Gadi>
local hard drives get picked up by our udev rule (except for vagrantc's who did REMOVABLE=1)
18:47
<ogra>
bryce removed even more stuff and fixed the pci id list
18:47
<Gadi>
or at least they should
18:47
<Q-FUNK>
I'm still waiting for nsc. sources hasn't been updated yet
18:47
<ogra>
it boils down to ten lines or so now
18:47
Gadi, which ubuntu inherited in ntrepid
18:47
i merges vagrants ltspfs
18:48
*merged
18:48
<Gadi>
ok
18:48
so , you are looking to unhack that hack?
18:48
:)
18:48
<ogra>
but i dont get why ntfs doesnt work for you
18:48
<Gadi>
not me, others I have helped in here
18:48
<ogra>
well, thats easy actually ... removing the REMOVABLE=1 is no prob
18:48
<Gadi>
it was usually not showing local ntfs partitions
18:49mighty-d has joined #ltsp
18:49
<ogra>
which it is supposed to do by default
18:49
<Gadi>
the deeper issue is having access to ALL files on a local hdd
18:49* ogra curses that nobody filed a bug for hardy about that
18:49
<ogra>
tha has to wait until 8.04.2 i guess
18:50
<mighty-d>
hi, im having great problems with gutsy and ltsp 5.0 , i had a slackware and ltsp 4.2 and it worked great, but ltsp 5.0 on gutsy goes too slow, my thin clients are celeron 800 and 256 ram
18:50
<ogra>
yes some people seem to ant that
18:50
<Gadi>
ithe problem was more one of: joe user shouldnt be able to boot as a thin client and then access the Admin files on his Windows partition
18:50
<ogra>
right, iirc the original debian bug added an lts.conf option
18:51
so you could switch local access on and off
18:51
i guess i'll reintroduce that in ubuntu
18:51
<Gadi>
well, one of the other flags I added personally, was to exclude a bus type
18:51
<ogra>
and look into why ntfs broke
18:51
<Gadi>
so, for example
18:51
I want all internal devices to work but not USB
18:51
LOCALDEV_EXCLUDE = usb
18:52
excludes all usb local devices
18:52* ogra wouldnt have guessed :P
18:52staffencasa has quit IRC
18:52
<Gadi>
(keyed off of udevinfo stuff)
18:52
<ogra>
hmm
18:52
cant you do that insde the rule ?
18:52
<Gadi>
it is
18:52
<ogra>
whithout separately calling udevadm ?
18:52
<Gadi>
its inside ltspfs_entry
18:53
I dont
18:53
<ogra>
oh, you just use the env vars ?
18:53
<Gadi>
I meant to say I use the BUS = usb
18:53
right
18:53
<ogra>
ah
18:53
<Gadi>
exactly
18:53
<ogra>
good
18:53
because you said udevinfo :)
18:53
<Gadi>
yeah, well do what I say not what I do
18:53
:)
18:53
or something
18:54
<ogra>
heh
18:54
<Gadi>
but, lots of nice switches to be had by calling ltsp_config in a udev program
18:54
:)
18:55
did you guys ever implement the "lets call getltscfg once" idea?
18:56
<mighty-d>
hello, would you help me?, ive been googling around and messing with this, but i havent come with a solution... im even considering to downgrade to 4.2
18:56
<Gadi>
mighty-d: sorry - we talk too much
18:56
whats up?
18:56
mostly me (wrt talking too much)
18:56
<mighty-d>
Gadi, thanks
18:56* ogra actually needs to leave
18:57
<ogra>
2am ...
18:57
<Gadi>
heh
18:57
mighty-d: LTSP5 too slow?
18:57
slow to boot or slow once you are logged in?
18:57
<mighty-d>
gadi, i had this ltsp on slackware running 4.2, and ltsp 5.0 clients are too slow, not only at boot but at run time
18:57
<ogra>
LDM_DIRECTX might help
18:57
<Gadi>
^^^
18:57
<ogra>
even though gutsy pretty much ucks
18:57
*sucks
18:57
i'D recommend hardy
18:57
<Gadi>
heh
18:58
<mighty-d>
ogra, so do you recommend to switch to hardy?
18:58
<ogra>
it has the same ltsp but tons of extra fixed
18:58
*fixes
18:58* vagrantc has almost never seen ogra recommend anything other than the current release
18:58* Gadi agrees
18:58
<ogra>
hardy is long term support release so the focus was on stabilty and steadiness of the code
18:58
<Gadi>
heh
18:59
oh... he was serious
18:59
:P
18:59
<mighty-d>
lol
18:59
<ogra>
no nifty extra features but much time invested into stabilization
18:59
Gadi, yes i was
18:59
<Gadi>
"stable is a state of mind, sir"
18:59
I was told that by the guy who ported Unreal Tournament to Linux
18:59
when I asked him
18:59* Gadi has to go put the kids to bed
19:00
<Gadi>
l8r
19:00
<ogra>
hardy ltsp is far beyond the gutsy version and unlike "normal" ubuntu releases will see regular fixes and updates
19:00
<mighty-d>
thanks gadi,
19:00
<ogra>
anyway
19:00* ogra &
19:00
<mighty-d>
bye ogra... thanks
19:12mighty-d has quit IRC
19:13aruiz has joined #ltsp
19:21Gadi has left #ltsp
19:25vagrantc has quit IRC
19:29Q-FUNK has quit IRC
19:30Q-FUNK has joined #ltsp
19:32Q-FUNK has quit IRC
19:36rjune has quit IRC
19:39juanpaul has quit IRC
19:54mccann has quit IRC
20:04RoninBaka has quit IRC
20:06savetheWorld has quit IRC
20:06aruiz has quit IRC
20:11rjune has joined #ltsp
20:12gregbrady has joined #ltsp
20:37Rucal has joined #LTSP
20:37Rucal has left #LTSP
21:01Lumiere has quit IRC
21:19johnny2 has left #ltsp
21:26Rucal has joined #LTSP
21:30Ruca1 has joined #LTSP
22:06cyberorg has quit IRC
22:12savetheWorld has joined #ltsp
22:17gregbrady has quit IRC
22:51RoninBaka has joined #ltsp
22:54indradg has joined #ltsp
23:06Ronin1234 has joined #ltsp
23:12cyberorg has joined #ltsp
23:18RoninBaka has quit IRC
23:20vagrantc has joined #ltsp
23:21captain_magnus has quit IRC
23:22captain_magnus has joined #ltsp
23:40Egyptian[Home1 has joined #ltsp
23:45Nuba1 has quit IRC
23:47Nubae has joined #ltsp