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


Channel log from 13 March 2008   (all times are UTC)

00:01
<johnny>
ok.. here's another common use case :)
00:02
amd64 server, x86 clients
00:02
relying on the server's portage config isn't useful for me atm
00:02
either
00:04
even if i was x86
00:05
DEBUG: Loading plugins in MODE=after-install:
00:06
so.. that'll tell you where i'm at now :)
00:15
<dberkholz>
johnny: that's what PORTAGE_CONFIGROOT is for
00:17
<johnny>
and the advantage over just chroot emerge is?
00:18
<dberkholz>
johnny: you don't need portage installed on the clients
00:18
johnny: plus chroots suck imho
00:18
being able to do stuff from outside the chroot is much more powerful
00:19
<johnny>
hmm..ok.. guess gotta reevaluate using quickstart then :(
00:20
<dberkholz>
nah
00:20
<johnny>
it already installs portage
00:20
<dberkholz>
i'm just saying that as a general guideline
00:20
familar with the --unmerge option? =P
00:21
<johnny>
sure
00:21
and the tree?
00:22
<dberkholz>
drop portage from the client at the same time you're cleaning out build deps
00:22
bind mount, if quickstart needs to run emerge inside the chroot
00:22
<johnny>
aha.. good idea..
00:22
at least the tree is much smaller these days..
00:39elisboa__ has joined #ltsp
00:51elisboa_ has quit IRC
01:52moquist has quit IRC
01:57basanta has joined #ltsp
02:07sepski has joined #ltsp
02:08
<johnny>
dberkholz, emerge --update --world is not a good option atm, due totally out of date stage tarball
02:08
issues with com_err and ss
02:15sepski has quit IRC
02:18deavid has joined #ltsp
02:57chupa has quit IRC
03:10hari has left #ltsp
03:13chupa has joined #ltsp
03:24clemyeats has joined #ltsp
03:25
<clemyeats>
Hi, is it possible to make an ltsp client image for a full X11 working environment? (I mean, the client would simply get the image from the server, and from there on run without the server).
03:48
<daduke>
clemyeats: err wouldn't a plain vanilla tftp/nfs boot do the trick? What do you need LTSP for in this case?
03:48
<clemyeats>
daduke: I'm not sure I do..
03:48
<daduke>
clemyeats: well I don't think you do ;)
03:48
<clemyeats>
daduke: ideally I'd like to use a liveCD ISO and have it served by the server to the network clients.
03:49bobby_C has joined #ltsp
03:50chupa has quit IRC
03:50
<daduke>
clemyeats: I've never tried it, but you should be able to put the contents of said ISO on an NFS server and then tftp boot its kernel. google for knoppix netboot or something perhaps?
03:50chupa has joined #ltsp
03:51
<clemyeats>
daduke: I will. Thanks a lot for this. Things are a bit clearer for me now. Thanks.
03:51
<daduke>
clemyeats: you're welcome.
03:51
<chupa>
oops, i got lost
03:55
<tarzeau>
HELP!
03:57
<daduke>
tarzeau: was?
03:57subir has joined #ltsp
04:04Nubae has joined #ltsp
04:12chupa has quit IRC
04:14chupa has joined #ltsp
04:47chupa has quit IRC
04:52Q-FUNK has joined #ltsp
04:59
<ogra_cmpc>
clemyeats, https://wiki.ubuntu.com/LiveCDShareThisCD
05:00
<clemyeats>
ogra_cmpc: perfect, thanks a lot for this!
05:05chupa has joined #ltsp
05:06
<cyberorg>
ogra_cmpc, was local hard disk auto mount changed after this? http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=432024
05:22Q-FUNK has quit IRC
05:26
<ogra_cmpc>
cyberorg, sisne loocal HD mounting was never tested i cant tell you if it changed :)
05:26
*since local
05:27
the code did change to prevent them completely though
05:28bobby_C has quit IRC
05:32basanta has quit IRC
05:44TiagoAT has joined #ltsp
05:46
<TiagoAT>
Alguém que possa responder?
05:46TiagoAT has quit IRC
05:47moquist has joined #ltsp
06:01chupa has quit IRC
06:02Q-FUNK has joined #ltsp
06:03chupa has joined #ltsp
06:10Q-FUNK has quit IRC
06:37steph_ has joined #ltsp
06:50
<vlt>
Hi, I received a spam mail today: "My name is lilian. I am a nice young lady ... Because I am tired of loneliness ... and saw your profile at (http://www.ltsp.org) and accepted you believing we match each other for love and friendship. ..."
06:51
lilian? Are you here?
06:51
;)
06:54
<ogra_cmpc>
haha
07:01Guaraldo has joined #ltsp
07:01daduke is now known as lilian
07:01Blinny has joined #ltsp
07:04subir has quit IRC
07:07edistar has joined #ltsp
07:08gentgeen__ has joined #ltsp
07:17Q-FUNK has joined #ltsp
07:18edistar_ has quit IRC
07:24deavid has quit IRC
07:25deavid has joined #ltsp
07:38
<cyberorg>
ogra_cmpc, prevent it? now i want to enable it :(
07:39
<ogra_cmpc>
did you find out what hogs the devices ?
07:39
(doesnt help much to poke on ltspfs if you didnt solve the general prob first)
07:39
<cyberorg>
ogra, nope, since usb plugging works, and the device nodes are created, udev is working, i removed all the other storage rules, still no go
07:40slidesinger has joined #ltsp
07:40steph_ has quit IRC
07:40
<ogra_cmpc>
well, get that sorted first then we can look at ltspfs
07:41
<cyberorg>
ogra_cmpc, get sorted what? devices are created, udev works, onlything that does mounting is 88-ltsp.rules, where else do i look?
07:42
<ogra_cmpc>
so you dont have any /sdv/sda anymore now ?
07:42
*/dev/sda
07:43
<cyberorg>
ogra, i have /dev/sd*
07:43
<ogra_cmpc>
/dev/sda, the local disk is what i'm talking about
07:43
<cyberorg>
ogra_cmpc, yes they are there
07:43
<ogra_cmpc>
make sure udev doesnt hog (create them)
07:43
than we can look at adding them to the ltspfs rules
07:44
but first you must make sure they are not claimed already
07:44
<cyberorg>
ogra_cmpc, they are not claimed, i made sure of that, removed all rules about storage except ltsp.rules
07:45
<ogra_cmpc>
if they are not claimed they wont exist
07:45
which they apparently do
07:45
<cyberorg>
ogra_cmpc, would they be created and lts rule ignores them because they are local disks?
07:45
<ogra_cmpc>
no
07:45
<cyberorg>
the bug i posted wanted something like that enforced
07:46
<ogra_cmpc>
if they are created they are owned by soemthing already
07:46
<cyberorg>
http://bugs.skolelinux.no/show_bug.cgi?id=1209
07:46
<ogra_cmpc>
right, thats the current state
07:46
which means if you really have cleaned everything, firing the ltspfs rule wouldnt create them
07:47
if they *exist at all* you need to get that sorted first
07:47
<cyberorg>
ogra_cmpc, ok, so ltsp rules wont create it at all, right, i'll dig deeper
07:48
<ogra_cmpc>
with the autoremove=1 it wont touch and non removable devices
07:48
s/and/any/
07:49
<cyberorg>
where is that configured?
07:50soneyka has quit IRC
07:50
<ogra_cmpc>
in the rules file yo pointed to several times and which ios quoted in the bug
07:52
<cyberorg>
ogra, grep autoremove * -R in all ltsp*-trunk ldm-trunk did not return anything
07:54krishna_ has joined #ltsp
07:58
<ogra_cmpc>
ltspfs
07:58
ihttp://bugs.skolelinux.no/show_bug.cgi?id=1209
07:58* ogra_cmpc sighs
07:59
<ogra_cmpc>
the thing you point me to since i woke up ...
07:59Nubae has quit IRC
08:00mhterres has joined #ltsp
08:00Q-FUNK has quit IRC
08:00
<cyberorg>
good morning :)
08:05deavid has quit IRC
08:06soneyka has joined #ltsp
08:21Gadi has joined #ltsp
08:25soneyka has quit IRC
08:41soneyka has joined #ltsp
08:44Q-FUNK has joined #ltsp
08:48joebaker has joined #ltsp
08:56Q-FUNK has quit IRC
08:56gonzaloaf has joined #ltsp
09:01tux_440volt has joined #ltsp
09:07mischko has joined #ltsp
09:12mischko has quit IRC
09:19nicoAMG has joined #ltsp
09:22cyberorg has quit IRC
09:37tux_440volt has quit IRC
09:48vagrantc has joined #ltsp
10:00mccann has joined #ltsp
10:04abadger1999 has quit IRC
10:20cyberorg has joined #ltsp
10:24abadger1999 has joined #ltsp
10:24alekibango has quit IRC
10:25alekibango has joined #ltsp
10:26mccann has quit IRC
10:28mccann has joined #ltsp
10:28mikkel has joined #ltsp
10:28staffencasa has joined #ltsp
10:37krishna_ has left #ltsp
10:44abadger1999 has quit IRC
10:56clemyeats has quit IRC
11:00mccann_ has joined #ltsp
11:03
<Gadi>
ogra_cmpc: ping
11:04
<ogra_cmpc>
Gadi, ?
11:04
<Gadi>
hey - would you happen to know if in gutsy, the CONSOLE_KEYMAP list is the same as the XKBLAYOUT?
11:04
or how to retrieve the full list of options
11:05
for CONSOLE_KEYMAP
11:05
is it just: /usr/share/keymaps/i386/qwerty/
11:06joebaker has quit IRC
11:08abadger1999 has joined #ltsp
11:10
<ogra_cmpc>
probably from /var/lib/dpkg/info/console-setup.config ?
11:10
<Gadi>
hmm... I found online a pseudo-man page for ckbcomp
11:10
looks like it is the same as XKBLAYOUT
11:10
as it computes from XKB
11:12
YAY! more hard-coded keymaps! :)
11:13
<ogra_cmpc>
??
11:13mccann has quit IRC
11:13* Gadi reads /var/lib/dpkg/info/console-setup.config
11:13
<ogra_cmpc>
thats surely generated during package build
11:13
not hardcoded
11:13
<Gadi>
ah, ok
11:15
<ogra_cmpc>
oh, fun ...
11:15
discovering that you have a typo at the end of a script that takes 2h torun is always a pleasure
11:16
grmbl
11:16
<Gadi>
should have run with --fix-typo
11:16
<ogra_cmpc>
damned, why didnt you say that earlier
11:16
now its already running agin
11:18* Gadi wonders if it is safe to set CONSOLE_KEYMAP = XKBLAYOUT
11:18
<ogra_cmpc>
i think the config file says that
11:18
<Gadi>
the code does: ckbcomp -model pc105 "$CONSOLE_KEYMAP"
11:19
that config file is not very readable
11:20
<ogra_cmpc>
?
11:21
/etc/default/console-tools isnt readable ?
11:21
<Gadi>
oh, no - I meant: /var/lib/dpkg/info/console-setup.config
11:21
<ogra_cmpc>
oh, you talk about the debconf file
11:21
well, thats a package configuration options list for debconf
11:21
usually nothing users should touch :)
11:22
<Gadi>
right
11:22
I just want to make sure that what console-setup uses XKBLAYOUT for is the same as when we do:
11:22
ckbcomp -model pc105 "$CONSOLE_KEYMAP"
11:23* Gadi admits this is a dev question - and why should devs get readable docs! :P
11:23
<ogra_cmpc>
# The following variables describe your keyboard and can have the same
11:23
# values as the XkbModel, XkbLayout, XkbVariant and XkbOptions options
11:23
# in /etc/X11/xorg.conf.
11:23timlegge has joined #ltsp
11:23
<Gadi>
right
11:23
but are they used to do:
11:23
<ogra_cmpc>
and there is no other keymap setting anywhere in that file
11:24
<Gadi>
ckbcomp -model pc105 "$XKBLAYOUT"
11:24
?
11:24* Gadi has to assume yes
11:24
<ogra_cmpc>
well
11:24* Gadi further assumes that ogra never wrote the ltsp code: ckbcomp -model pc105 "$CONSOLE_KEYMAP"
11:24
<ogra_cmpc>
whats actually used is /etc/console-setup/boottime.kmap.gz
11:25
unless you trigger siomething by hand
11:25
\and that is set on package install
11:25
<Gadi>
am I the only one here working on LTSP? :)
11:25
in ltsp-client-setup, we run: ckbcomp -model pc105 "$CONSOLE_KEYMAP"
11:26abadger1999 has quit IRC
11:26
<Gadi>
I dont care what Ubuntu PCs do by default
11:26
<ogra_cmpc>
huh ?
11:26
<Gadi>
I just want to understand what *we* do
11:26
<ogra_cmpc>
where qand when was that added ?
11:26
<Gadi>
look in the code
11:26
thats been there since the beginning of time
11:26
must have ben mdz
11:26
<ogra_cmpc>
beta freeze today
11:27* ogra_cmpc has no time to dig deep into ltsp atm
11:27
<Gadi>
ok
11:27
<ogra_cmpc>
in any case that needs to go
11:27
<Gadi>
why?
11:27
<ogra_cmpc>
ltsp-build-client runs console-setup configuration during build
11:27
we shouldnt5 muck around manually there
11:28
<Gadi>
unless you want a different keymap than the default locale
11:28
<ogra_cmpc>
if we need to there needs to be a proper preseeding mechanism and a dpkg-reconfigure in place
11:29
if you want a different keymap add that code :)
11:29
<Gadi>
lol
11:29mccann_ has quit IRC
11:29* Gadi moves on to more important things...
11:29
<ogra_cmpc>
its the only place where it should be set
11:29
<Gadi>
as many have most likely done in the past
11:29chupa has quit IRC
11:30
<ogra_cmpc>
since it will care for all dependent changes that habve to be made if there are any
11:30
and will not break on upgrades
11:31mccann_ has joined #ltsp
11:32
<ogra_cmpc>
Gadi, put up proper preseeding with debconf-communicate to set trhe debconf values for keymap and model and run dpkg-reconfigure console-setup >/dev/null 2>&1
11:33
<johnny>
dpkg-reconfigure classmate-pc
11:33
done
11:33chupa has joined #ltsp
11:33
<johnny>
ok ogra, finished your work for ya :)
11:33
<ogra_cmpc>
haha
11:33spectra has joined #ltsp
11:35
<johnny>
hmm... soon it will time time to get back to the kernel..
11:35
be time*
11:36
if i do that, and have a couple ebuilds, gentoo support will be on it's way..
11:36
<ogra_cmpc>
cool
11:36
<johnny>
it is really helping me understand some inner bits of ltsp
11:44daduke_ has joined #ltsp
11:46lns has joined #ltsp
11:46
<lns>
Hey all
11:47abadger1999 has joined #ltsp
11:54abadger1999 has quit IRC
11:54abadger1999 has joined #ltsp
12:12abadger1999 has quit IRC
12:12abadger1999 has joined #ltsp
12:15highvoltage has joined #ltsp
12:15
<highvoltage>
where can I find sbalneav's fancy documentation again?
12:19
<ogra>
highvoltage, apt-get install edubuntu-docs
12:19
<highvoltage>
ogra: ooh, thanks!
12:19
<ogra>
its all in the edubuntu handbook
12:19
(and only there sadly)
12:20
<highvoltage>
hmm, where do I find it again? I looked in file:///usr/share/doc/edubuntu-docs/ but I only find the copyright file, etc there
12:20
<ogra>
in yelp :)
12:20
startpage, right side at the bottom
12:22
<highvoltage>
wow, this is nice documentation. can't see where it mentions how to disable ssh encryption though, but I'll keep digging a bit more there...
12:23
<vagrantc>
LDM_DIRECTX=true
12:24
<highvoltage>
thanks vagrantc
12:26
<vagrantc>
i think gutsy was the first ubuntu release to support it
12:27
<ogra_cmpc>
right
12:27
and i'm niot sure scott updated the docs
12:27
<dberkholz>
johnny: keep the emerge world regardless, doesn't matter how out of date it is, it's what we'll have and what all the users will use (same thing they'd do on any new install)
12:27
<johnny>
i plan to keep it
12:27
i just say that it is totally broken :)
12:30
!!! Multiple versions within a single package slot have been
12:30
!!! pulled into the dependency graph:
12:30
<ltspbot>
johnny: Error: "!!" is not a valid command.
12:30
johnny: Error: "!!" is not a valid command.
12:30
<vagrantc>
dberkholz: did we ever figure out if you have admin privledges on launchpad ?
12:31
<ogra_cmpc>
vagrantc, just make him one
12:31
<dberkholz>
vagrantc: i don't remember exactly how you added me
12:31
<johnny>
dberkholz, i'm assuming that a 2008.1 stage tarball will fix it
12:31* vagrantc casts the proper launchpad incantation
12:31
<ogra_cmpc>
:)
12:31
<dberkholz>
johnny: grab tarballs from funtoo.org to test if you want
12:31
<vagrantc>
ogra_cmpc: had it been clear how to do so, there would be no doubt weather it had been done.
12:32
<dberkholz>
drobbins has been posting weekly stages
12:32
<johnny>
oh.. daniel robbins has some?
12:32
aha
12:32
thanks!
12:32
i would be willing to build one myself, if i knew the proper way
12:32
<ogra_cmpc>
vagrantc, https://launchpad.net/~ltsp-upstream/+members
12:32
<johnny>
but using somebody else's is great too
12:33
<dberkholz>
johnny: you use catalyst
12:33
<ogra_cmpc>
click on the pen next to donnie
12:33
<johnny>
sure, i have catalyst, but then what? :) i'd like to use the same method as releng, of which i do have the svn repo, but i don't know what to do after
12:34
<vagrantc>
ogra_cmpc: unless i'm missing something, i don't have privs to make dberkholz an admin
12:34
<dberkholz>
johnny: run catalyst =)
12:34
<ogra_cmpc>
vagrantc, what do you see on the detailed page (after clicking teh pen)
12:35
<dberkholz>
if you have the spec files and a portage snapshot, i don't recall it being a difficult process
12:35
<vagrantc>
ogra_cmpc: Administrator: No
12:35
<ogra_cmpc>
there should be a radiobutton for admin next to the "member since"
12:35
<vagrantc>
ogra_cmpc: no radiobutton
12:35
<johnny>
hmm.. no stage3.. only 1 and 2..
12:35
<ogra_cmpc>
hmm, weird
12:36
<johnny>
oh.. the previous ones have stage3
12:36
just not the current ne
12:36
<vagrantc>
ogra_cmpc: i'm wondering if "team owner" has more privs than administrators ?
12:36
<ogra_cmpc>
apparently
12:37
<johnny>
i'm happy using somebody else's tho,since that means i don't have to waste cpu time building one myself
12:37
i need another box i guess
12:37
<ogra_cmpc>
i wonder if i can make ltsp-upstream own ltsp-upstream
12:37
<dberkholz>
bbl
12:37mccann has joined #ltsp
12:37
<ogra_cmpc>
The person/team named 'ltsp-upstream' is not a valid owner
12:37
gah
12:37
<johnny>
gah! :(
12:38
he's on a good server.. 800K down..
12:38
<ogra_cmpc>
vagrantc, try now
12:38
Summary should have changed for you as well
12:39
to Restricted Team
12:39
apparently "moderated team" only allows the owner to approve admins
12:40
<vagrantc>
ogra_cmpc: no luck
12:40
ogra_cmpc: i can remove my own administrator status :)
12:41
<ogra_cmpc>
well, it was moderated which apparently was the wrong thing ... but now its restricted
12:41
Restricted Team: New members can only be added by one of the team's administrators.
12:41abadger1999 has quit IRC
12:41
<ogra_cmpc>
that should give you the right
12:42
well, i'll do it for now and try to not forget about asking in #launchpad
12:43
<vagrantc>
ogra_cmpc: yeah, i see the restricted team stuff now
12:43
but still no way to make someone an admin ..
12:43
<ogra_cmpc>
but you cant change
12:43
<johnny>
dberkholz, running again :)
12:43
<vagrantc>
ogra_cmpc: i think we've found a bug.
12:44
ogra_cmpc: it does give me a radiobutton to disable your admin status
12:44
<ogra_cmpc>
dberkholz, are you promising to use your powers with care and not add people to the team that fuck up everything ?
12:44
<johnny>
ogra_cmpc, i think you can trust him
12:44
he does very well with gentoo :)
12:45
<ogra_cmpc>
vagrantc, well, it should then also give you one for dberkholz
12:45
<vagrantc>
ogra_cmpc: yes. should and does are two different things in this case.
12:45
<ogra_cmpc>
funny
12:46
<vagrantc>
i can take away admins, but not add them :)
12:46* ogra_cmpc makes dberkholz and admin
12:46
<ogra_cmpc>
*an
12:46
vagrantc, can you take away his status ?
12:47
<johnny>
is it possible to only allow commit access to certain branches?
12:47
or creation of branches, but not addition?
12:47
err not commit to certain branches
12:47
<vagrantc>
ogra_cmpc: yup.
12:47
<ogra_cmpc>
something at least
12:47mccann_ has quit IRC
12:47
<johnny>
here we go again..
12:48
<ogra_cmpc>
johnny, you can create a branch and allow access for people or teams, yes
12:48
<johnny>
so, who here uses ltsp on multiple distros?
12:49* ogra_cmpc crosses fingers watching tonight scmpc image booting .... (belated after we fucked up glibc and had no archive the whiole day)
12:49soneyka has quit IRC
12:52
<johnny>
hmm..
12:52soneyka has joined #ltsp
12:52
<johnny>
dberkholz, when doing that step, i get
12:53
emerge -uD world with PORTAGE_CONFIGROOT
12:53
scanelf.c:1: error: CPU you selected does not support x86-64 instruction set
12:54
i'm having trouble guessing that it's even using my config in any case, as i'm actually getting a message like
12:54
Overlay eclasses override eclasses from PORTDIR
12:54
since i'm using the gnome overlay
12:55
shouldn't that be ignored since i'm using PORTAGE_CONFIGROOT ?
12:56indradg has joined #ltsp
12:57
<johnny>
ogra_cmpc, how common is the usage of 64bit servers with 32bit clients?
12:58
sounded pretty common from the times i've been around here
12:59
<ogra_cmpc>
yeah, people do crazy stuff
12:59
sadly its quite common
13:00
<johnny>
what's the problem with it?
13:00
other than flash..
13:00
<ogra_cmpc>
flash
13:00
<johnny>
other than flash :)
13:01
<ogra_cmpc>
thats usually enough to generate some weeks of support in here
13:01
<johnny>
sure, but i dont' even have flash
13:01
at the store
13:01
it is disabled
13:02
so would you consider it a problem in any other case other than flash?
13:02
<ogra_cmpc>
java probably
13:03
but that should work if you use the 64bit ff wiuth the 64bit icedtea
13:03
<vagrantc>
there's still some hard-coding in some of the ltsp tools that doesn't behave well with cross-arch stuff
13:03daduke_ has quit IRC
13:03
<ogra_cmpc>
the prob with flash is that people use the 32bit version on amd64
13:03
<warren>
My 64bit server defaults to 32bit clients
13:04
<ogra_cmpc>
and dont realize that ff needs far more than just some libs if you want all features to work right
13:04
warren, i meant the 32bit version of ff on a 64bit env
13:04
<johnny>
i'm using nspluginwrapepr on my own box
13:04
nspluginwrapper
13:05
<warren>
make sure you use a very recent nspluginwrapper
13:05
they fixed a critical glib threading bug that latest flash needs
13:05
<johnny>
no problems so far..
13:05
<warren>
otherwise it crashes more often than usual on multi-core cpus
13:05
took me WEEKS to figure that out
13:06
<Blinny>
What kind of performance hit do you take using 32-bit w/ PAE-enabled kernel though? I've not yet had time to do tests.
13:06
<warren>
Blinny, 5-20% depending on the hardware and what software you are using. There are benefits to using x86_64 kernel with 32bit userspace.
13:07
<ogra_cmpc>
thats what ubuntu does by default since gutsy
13:07
we dont have the amd64 kernel flavour anymore
13:07
<Blinny>
ogra_cmpc: 64-bit kernel and 32-bit userspace?
13:07
<ogra_cmpc>
right
13:07
<Blinny>
Woah. I didn't know that.
13:07
<ogra_cmpc>
ubuntu only has a -generic kernel
13:08
<warren>
ogra_cmpc, eh? one kernel boots either 32bit or 64bit?
13:08
<ogra_cmpc>
well, and some specific other ones, but no arch separation anymore
13:08
warren, no, the same kernel installed on either environment
13:09
s/kernel/kernel binary/
13:09
<Blinny>
ogra_cmpc: So, come Hardy final release time, one can just install the i386 variant and get a 64-bit-enabled kernel out-of-the-box, sans PAE
13:09
<warren>
ogra_cmpc, I don't understand
13:10
<ogra_cmpc>
Blinny, right, the kernel binary package is generalized
13:11
warren, you have the same binary ... no matter if you use the amd64 cd or the i386 cd for installing your system
13:11
<Blinny>
ogra_cmpc: Coo. Thank you.
13:11
<warren>
ogra_cmpc, vmlinuz has both 32bit and 64bit kernel?
13:11
or separate files?
13:11mccann has quit IRC
13:11
<ogra_cmpc>
vmlinuz ids the same on both arches
13:12mccann has joined #ltsp
13:12
<ogra_cmpc>
same binary
13:12
dont ask me about details i'm not in the kernel team, thats how it was explained to me by the kernel guys
13:13
i just know that we have about five different kernel flavours before and we now have one
13:13
<warren>
If you install Fedora's kernel-2.6.24.3-12.fc8 on i386, you get i386 kernel. If you install the same version number on x86_64, you get x86_64 kernel.
13:14
we have one kernel RPM per architecture. 32bit, 64bit, ppc, ppc64, s390, s390x, ia64 (and soon mips, arm, sparc64)
13:14
<ogra_cmpc>
we still have different arches
13:14
but no differentiation beytween 32 and 64 bit anymore
13:15
not sure how that is with ppc/ppc64
13:17
<vagrantc>
ubuntu doesn't even support powerpc anymore, no?
13:19
<ogra_cmpc>
community supported
13:19
*canonical* doesnt support ppc anymore
13:19
same goes for sparc
13:24
vagrantc, lpia might be added with hardy, not sure
13:24
<vagrantc>
ogra_cmpc: what's that?
13:24
<ogra_cmpc>
intel menlo
13:24
mobile cpu
13:24
<vagrantc>
ah.
13:25
i only have some guage of interesting old and likely obsolete stuff.
13:25
<ogra_cmpc>
technically we started supporting it at end of gutsy ... but i'm not sure about the canonical support status, might become official
13:25
and there were discussions about the cell arch
13:25
(playstation)
13:31elisboa_ has joined #ltsp
13:31tux_440volt has joined #ltsp
13:32
<warren>
PS3 Linux is fine for compute node with highly specialized code
13:32
but otherwise useless especially for UI's
13:32elisboa has quit IRC
13:32mikkel has quit IRC
13:35mikesh has joined #ltsp
13:35
<ogra_cmpc>
warren, depends if sony gives you the specs and how much evilness you bear, i know they offer an nda for devs
13:35
<warren>
ogra_cmpc, you know my tolerance level for evil
13:35
<ogra_cmpc>
so having a cd that enables devs to work on teh platform might be helpful for them
13:36
them == sony
13:36
<mikesh>
hello, i have problem with lts.conf, XF86CONFIG_FILE doesnt work, is there some bug or.. what?
13:38
<Gadi>
mikesh: did you set it to the full /path/to/file
13:39
?
13:39
<ogra_cmpc>
warren, well, my tolerance level corellates with my social sense ... there are situations where i think its ok to accept evil ... else i wouldnt work on the cmpc to get a laptop to every child ...
13:40
<mikesh>
Gadi: no, just /etc/X11/my.xorg.conf (full path is /opt/ltsp/i383/etc/X11/my.xorg.conf)
13:40
<Gadi>
that is correct - what you did
13:40
did you update the imag?
13:40
<warren>
ogra_cmpc, that's otherwise known as "flexible morals" =)
13:40
<ogra_cmpc>
mikesh, distro/ltsp release ?
13:40
<Gadi>
*image?
13:41
<mikesh>
ogra_cmpc: debian lenny, ltsp from apt, i dont know version
13:41
<ogra_cmpc>
warren, no, hat is known as social sense, i find the bigger piocture more importwant than my own ideology so if i can get the kids a latop by renting my butt out to intel i do that
13:41
<mikesh>
Gadi: yes
13:41
<Gadi>
ah, then nm the image comment
13:41
:)
13:41sepski has joined #ltsp
13:41
<Gadi>
try changing to: X_CONF
13:41
both *should* work but, the one you use is deprecated
13:45
<mikesh>
Gadi: thats it, thanks a lot
13:45
<warren>
ogra_cmpc, "I think NVidia 3D is important for social reasons, thus I'm willing to violate copyrights to ship it in my distro?"
13:46
<ogra_cmpc>
no, thats different
13:46
<Gadi>
np
13:46
<ogra_cmpc>
and you can include it in a way that it doesnt infringe coopyrights
13:46
<warren>
ogra_cmpc, that is a legal gray area
13:46
<ogra_cmpc>
well
13:46* laga gets out the popcorn
13:46
<ogra_cmpc>
anyway the reasons are different
13:46
<warren>
anyway neither of us will win this discussion.
13:47
I know what you are referring to in Sony
13:47
is different
13:47
In Sony's case you could be under NDA and develop
13:47
(as long you aren't doing GPLv3)
13:47
<ogra_cmpc>
laga, sadly both nicks (waldorf *and* stadler) are taken already :)
13:48
well, in caseof nvidia its simply "enable all users to use their hw"
13:48
as long as the area is grey thats fine ... if you step into clear inllegality its not imho
13:48
<warren>
ogra_cmpc, unless you are using an older nvidia card where they dropped you
13:49
<ogra_cmpc>
well, but then you are screwed anyway since its unlikely to be supported by anythig else than vesa
13:50
<warren>
Which goes back to the other reasons why depending on evil is not good for you. =)
13:51
<ogra_cmpc>
i didnt say i would make anything depend on evil ... :)
13:51
but i'm willing to make compromises
13:52
if i can achieve bigger goals through that
13:53
<laga>
i think nvidia still supports riva tnt2 cards thru their legacy-driver.. ;)
13:53
<ogra_cmpc>
in ubuntu at least, yeah
13:54
but it wont work better than vesa anyway
13:54
so you can save the time for fiddling :)
13:54
<laga>
it'll still give you Xv, no?
13:54
<ogra_cmpc>
hmm
13:54
might be ...
13:54
but thats something only myth people care about :P
13:54
<laga>
i think so
13:55* ogra_cmpc hides
13:55
<laga>
i recently saw a bug report where one user said he was using a riva tnt2 card with the legacy driver with mythtv
13:55* laga slaps ogra_cmpc
13:55
<ogra_cmpc>
:)
13:56
<warren>
Open spec ATI cards are making real progress
13:56
<ogra_cmpc>
yeah, heard that
13:56
<warren>
kind of funny, avivo made more progress than the PAID radeonhd effort
13:56
<laga>
heh
13:56
<warren>
but ati is ahead of both
13:56
<ogra_cmpc>
heh
13:56
<laga>
it's a bit sad too see two independent teams working on it..
13:56
<ogra_cmpc>
well, i turned into an intel fan if it comes to graphics
13:57* laga is not gonna buy ATI until their free drivers support proper 3d.
13:57
<warren>
too bad there's no intel-based thin clients yet
13:57
laga, me too.
13:57
<laga>
yeah, intel is nice.
13:57
<warren>
laga, competition is good.
13:57
<ogra_cmpc>
warren, i heard there is some thinking about it ...
13:57
<warren>
ogra_cmpc, I hope so
13:58
<laga>
i own a radeon 9250 which i bought because of the free drivers.. mythtv in fullscreen mode is very slow with EXA, works with XAA. and interlaced modelines have been broken for a few months now. i'm not happy. :)
13:58
warren: true.
13:58
<warren>
had a Radeon 9200 but it broke
13:59
laga, the generation after Radeon 9200 works decent with 3D now
13:59
<ogra_cmpc>
i dont think i run any machine that uses something else than onboard graphica anymore
13:59
*graphics
13:59
<laga>
warren: yeah, r300. the 9200 is still r200 IIRC
14:01deavid has joined #ltsp
14:04mccann has quit IRC
14:05mccann has joined #ltsp
14:11tux_440volt has quit IRC
14:11indradg_ has joined #ltsp
14:23mccann has quit IRC
14:23mccann has joined #ltsp
14:27
<johnny>
dberkholz, you back yet?
14:29indradg has quit IRC
14:42Blinny has quit IRC
14:47Gadi has left #ltsp
14:49sepski has quit IRC
14:55bobby_C has joined #ltsp
14:57indradg__ has joined #ltsp
14:58mccann has quit IRC
14:59mccann has joined #ltsp
15:02bobby_C has quit IRC
15:07bobby_C has joined #ltsp
15:14abadger1991 has joined #ltsp
15:15indradg_ has quit IRC
15:17mccann has quit IRC
15:22
<warren>
Hmm, I can't think of a good way to redo mkdst and automatically find the spec file elsewhere.
15:22
any ideas
15:22
?
15:23
I guess you can add NAME.spec to .bzrignore, and symlink create a symlink from there to your spec file elsewhere.
15:23
you would only have to do it once
15:24
and it wouldn't end up in the tarball or checkin
15:24
Yeah, I'll do it that way, because any other way would require complicated code that is likely to break.
15:26* vagrantc thinks a bit
15:26
<vagrantc>
warren: you need/desire it to automatically detect which spec file to use?
15:26
<warren>
vagrantc, no
15:27* vagrantc doesn't see the huge problem with a commandline option
15:27
<warren>
vagrantc, I want to retain the ability to instantly build an RPM with any change in my source tree, without having the spec in the tree.
15:27
vagrantc, mkdst rpm ... if NAME.spec does not exist, fail with an explanation. If exists, do it.
15:27mccann has joined #ltsp
15:28
<warren>
NAME.spec will either be a copy from elsewhere that doesn't exist in this tree permanently or a symlink to elsewhere.
15:28
<vagrantc>
i guess i'm not going to be building rpm's ... so i shouldn't invest too much energy in it :)
15:28
<warren>
I think this makes everyone happier and even easier for me, because I don't have to maintain the spec in multiple places.
15:28
<vagrantc>
indeed.
15:29
<warren>
After I redo mkdst and make it functional, we can work on splitting out the commands into plugin files.
15:29
so you can add more files for arbitrary other commands
15:29
mkdst susebuildservice
15:29
mkdst whateverbuildservice
15:29
<vagrantc>
i'm still maintaining the init scripts in two places ... *sigh*
15:30
<warren>
vagrantc, and I'm removing mkdst-wrapper, if mkdst is installed from a package you really don't need it in the base of each source tree
15:30
<vagrantc>
warren: yeah.
15:30
<warren>
vagrantc, nobody responded against that in my proposal
15:30
k
15:30
<vagrantc>
warren: totally in support of that.
15:30
<warren>
ok, implementing...
15:30
<vagrantc>
even if you don't have it installed in $PATH, you can specify a full path to it
15:31
<warren>
yeah
15:31
vagrantc, any idea how to implement version number comparison in bash?
15:31* warren hates doing math in bash
15:32
<vagrantc>
you switching it to bash ?
15:32
<warren>
oh
15:32
vagrantc, I dunno, it depends if dash can do the math comparison or not
15:32
vagrantc, given that this is a DEVELOPMENT tool and not an initscript bash shouldn't matter right?
15:33
<vagrantc>
warren: yeah, i don't care if it requires bash, although if you're doing things that need bash ... may as well use some other language.
15:33
that's my philosophy, anyways.
15:33
<warren>
C!
15:33
<vagrantc>
with a few exceptions ...
15:33
warren: could just use "expr" i think.
15:34
<warren>
I'll figure it out...
15:34
<vagrantc>
warren: but version number comparisons tend to be really complicated
15:34
<warren>
I know =(
15:34
<vagrantc>
i mean, simple cases are fairly easy to handle ... but if you want something robust ... you have a lot of work ahead of you :)
15:35
warren: what sort of version comparisons are you envisioning?
15:35
<warren>
vagrantc, release.conf:REQ_MKDST=0.8
15:35
mkdst-1.0 > mkdst-0.8
15:36
mkdst-0.7 < mkdst-0.8
15:36
vagrantc, maybe we don't really need a minimum mkdst marker
15:36
<vagrantc>
oh, so you're basically in control of the version numbering and don't have to support insane versions
15:36
<warren>
vagrantc, yeah
15:36
current version is mkdst-0.5
15:36
<vagrantc>
that's not so bad
15:36* vagrantc tends to like X.Y.Z versions ... even if Z is usually .0
15:37
<warren>
We can move from 0.Y in the next version if you want.
15:37
<vagrantc>
not something i'd spend much energy on :)
15:38
<warren>
vagrantc, we still have to talk about what version numbers to give upstream ltsp/ltspfs/ldm
15:38
<vagrantc>
yeah.
15:39
<warren>
not easy, given that both you and ogra are shipping things that aren't in trunk right?
15:39
<vagrantc>
in ubuntu and debian, we're up to ltsp 5.0.40~ (~ means "a little less that") ... and ltspfs is i think 0.5.0~ ... ldm is crazy ... debian's basically 0.1~ and ubuntu's 5.0.40~
15:39
<warren>
vagrantc, an option might be to JUST TAG NUMBERS but don't call it a release. They are only milestones.
15:39
vagrantc, that would be terribly useful for me
15:40
<vagrantc>
warren: i believe everything we're shipping is basically in bzr, with a few small exceptions.
15:40
warren: but not tagged as yet.
15:40
<warren>
ltspfs and ldm, let's just tag them with a number, call it only a milestone and move on?
15:41
my packages would then have sane progressions instead of CONSTANTVERSION-date
15:41
And we really have to branch ldm
15:41
<vagrantc>
warren: i'm not willing to branch ldm until someone actually has a branch to merge and say "this is different and should be trunk"
15:42
<warren>
vagrantc, sadly, we have to convert the metadata format in order to support bzr tags.
15:42
<vagrantc>
really? gah.
15:42
<warren>
vagrantc, yes, none of the 3 have dirstate tags
15:42
<vagrantc>
bzr is making me grumpy. even though i like it.
15:42* warren last I looked.
15:42
<warren>
apparently bzr became a GNU project
15:43
<vagrantc>
laga: you're using aufs for ltsp related stuff ?
15:44
unionfs is borked in debian at the moment ...
15:44* vagrantc notes that it's been a long moment, from the looks of things
15:44
<vagrantc>
squashfs is borked, too.
15:44
and people wonder why i stick with NFS
15:45
<warren>
how's fuse-unionfs?
15:45
<vagrantc>
NBD+squashfs+unionfs ... that's 3 different layers that can break.
15:45
<warren>
vagrantc, squashfs too? ugh.
15:46
<vagrantc>
i haven't tried funionfs lately...
15:46
was difficult to get fuse working in the initramfs, though
15:46
<warren>
what sort of issues?
15:47
<vagrantc>
it was months ago, i don't really remember.
15:48
the thing i eventually had working was NBD+squashfs+tmpfs bind mounts
15:49
i'd like to try next is NBD(with cow)+ext2
15:49
maybe add tmpfs bind mounts for the regularly written stuff
15:51
<warren>
what kind of cow?
15:51slidesinger has left #ltsp
15:51
<warren>
tmpfs bind mounts for regularly written stuff because cow will constantly grow because it lacks whiteouts?
15:51
or collapsing
15:51
or whatever it is called
15:51
does unionfs collapse?
15:51
Create file, delete file, is memory space recovered?
15:52highvoltage has quit IRC
15:53
<johnny>
you're a cow
15:53
hehe
15:54
i'm having trouble choosing bzr for my own projects because of it
15:54
i thought mtn was complicated.. but i'm having serious trouble learning bzr
15:54
like.. wtf does bzr make a backup file when i revert :(
15:55
why do i have a choose a branch format, etc :(
15:55
lame lame lame :(
15:56
<vagrantc>
warren: NBD's copy-on-write support
15:56
<warren>
sweet, really simple changes makes RPM build support much smoother
15:56
mkdst rpm --> mkdst-0.5-1.fc9.20080313.16
15:56
normal rpmbuild --> mkdst-0.5-1.fc9
15:56
vagrantc, oh! forgot about that.
15:57
vagrantc, is that implemented in nbd client itself or the server do you know?
15:57
<vagrantc>
warren: server-side
15:57
warren: so could be really useful for low-memory clients
15:57
<warren>
I see.
15:57
Yeah
15:57
<johnny>
hmm.. now i have to figure out how to use this setarch thing properly..
15:58
or else figure out another way to force compiles into the chroot from amd64 server :(
15:58
prolly need other magicz
15:58
<warren>
Gentoo huh?
15:58* vagrantc thinks doing as much in the chroot as possible is wise
15:58
<johnny>
vagrantc, dberkholz disagrees :)
15:58
<warren>
vagrantc, so he'll have to install the gcc and toolchain in the chroot? nasty.
15:59
<laga>
vagrantc: i use aufs
15:59* vagrantc feels popular
16:00
<vagrantc>
johnny: well, i'll conceed to dberkholz on all things gentoo ...
16:00
<dberkholz>
aww, you're so sweet =)
16:00* johnny needs to come to portland :(
16:00
<johnny>
why does it have to be on the completely opposite side of the country :(
16:00
<dberkholz>
gentoo shouldn't have any trouble doing stuff from outside the chroot, portage fully supports installing elsewhere from /, and it allows all the build deps to stay out of the client
16:00abadger1991 has quit IRC
16:00
<johnny>
dberkholz, the problem is compiling for x86
16:01
<dberkholz>
johnny: you might use crossdev to build out an x86 toolchain
16:01
<johnny>
well i shouldn't need to
16:01
<dberkholz>
although i think just adding -m32 to cflags on amd64 should do it
16:01
<vagrantc>
but in general, i think doing things inside the chroot rather than server-side is a good thing ... as it allows you to support different environments in a clean way ...
16:01
<johnny>
there's something about ABI
16:01
<warren>
vagrantc, hmm, here's an issue with NBD cow
16:01
vagrantc, your image can't be squashfs
16:01
<dberkholz>
johnny: if you're using PORTAGE_CONFIGROOT=/opt/ltsp/i386 and you have the profile set up there properly, it should work
16:01
<johnny>
i can't tell if you can set it as an env variable
16:01
<dberkholz>
johnny: not saying it _does_ work, but it should
16:01
<vagrantc>
warren: my images can't use squashfs right now anyways :)
16:01
<dberkholz>
johnny: ABI comes from the profile
16:01
<johnny>
aha.. true
16:01
<dberkholz>
so should be a matter of getting profile for clients respected
16:02
<warren>
vagrantc, so you would export an ext3 image raw?
16:02
<vagrantc>
warren: ext2, but yeah
16:02
<johnny>
scanelf.c:1: error: CPU you selected does not support x86-64 instruction set
16:02
<warren>
vagrantc, the temporary COW file would have to be the same size as the original right? wont scale?
16:02
<johnny>
first update is to pax-utils
16:02
<vagrantc>
warren: i think it incrementally adds space as needed.
16:02
warren: i intend to experiment with it and i'll have a much better idea shortly.
16:03
<warren>
vagrantc, using a sparse file or something?
16:03
<dberkholz>
johnny: try searching around a bit (google etc) for "cross-compiling" amd64 to x86 on gentoo
16:03
<johnny>
been doing it
16:03
a toolchain should be unnecessary tho..
16:03
i mean.. a fully x86 toolchain that is :)
16:04
<dberkholz>
right i understand what you're saying
16:04
<johnny>
i'll figure it out
16:04
just talking it out with myself a bit
16:04* vagrantc wonders about starting off with cross-arch builds rather than just getting a basic build first
16:04
<dberkholz>
yes, i also wonder about this.
16:05
<johnny>
the build will work
16:05
minus the kernel
16:05
but that requires fixing genkrnel4..
16:05
<warren>
wow, already done with moving spec elsewhere.
16:06
<johnny>
i'm already near the last step of the build
16:06
<dberkholz>
johnny: you could certainly do http://www.gentoo.org/proj/en/base/amd64/howtos/chroot.xml
16:06
<johnny>
yes, i read that
16:07
i just wanted to do it the right way, since it is a very common use case
16:07
<dberkholz>
my approach is to first get the simplest case working, but think about what might need to change in more complex cases
16:07
as vagrantc was saying, it seems like you're already trying to start out with a more complex one.
16:07
<johnny>
does virtualbox work with 64bit yet?
16:07
otherwise i have no clients to test with anyways
16:08
<dberkholz>
is it keyworded for amd64?
16:08
<johnny>
it works in amd64
16:08
<warren>
johnny, don't need clients to test it =)
16:08
<johnny>
but last i heard, it won't let you boot into one
16:08
<dberkholz>
if virtualbox doesn't, qemu should
16:08
<johnny>
aha.. qemu.. forgot about that..
16:09mccann has quit IRC
16:09
<johnny>
how hard is that to setup? any good docs?
16:09
<warren>
johnny, the very minimal config to do qemu is in ltsp-trunk/vmclient
16:09mccann has joined #ltsp
16:09
<warren>
just tweak the stuff in ltsp-vmclient script
16:09
<johnny>
ok.. cool..
16:09
<warren>
I plan on generalizing that script more at some point
16:10
<johnny>
ok.. time to download another stage3..
16:10
<dberkholz>
virt-manager is pretty nice for playing w/ qemu
16:10
<warren>
johnny, only problem, if you don't have hardware virt then qemu is possibly too slow to use for this testing. You will want to use virtualbox or vmware instead.
16:10
<johnny>
oh.. i don't :(
16:10
<dberkholz>
warren: yeah it's pretty slow, but it's usable for basic tests...
16:10
<warren>
I have no idea how to PXE boot with virtualbox or vmware, but I know they can do it.
16:10
<johnny>
warren, it's very easy
16:10
too easy :)
16:11
2 clicks away or so
16:11
i use it to tst my ubuntu vm
16:11
err ubuntu ltsp server
16:12
my deployment at the store is all ubuntu
16:12
<vagrantc>
for thin client testing, qemu is useable ... slow, sure. but definitely useable to see that it boots and logs in
16:13
<johnny>
ok.. i'll get back on this tonight
16:13
<dberkholz>
btw johnny i requested a new overlay for ltsp
16:13
<johnny>
first i must go work in aforementioned store
16:13
how does that work?
16:13
do they do the hosting?
16:13
<dberkholz>
yes
16:13
and we can get you access once it's set pu
16:13
<johnny>
gonna make me learn git?
16:13
:(
16:14
so.. day to day.. i must use.. svn,bzr,mtn, and git :)
16:14
and cvs.. oops
16:14
5 systems in one day.. kinda tough :)
16:14
<dberkholz>
yeah, i need to finish my generic scm wrapper.
16:15
<johnny>
dberkholz, i'll show you some bzr eclasses, and let you pick the best one
16:15
unless you have one in mind already
16:15
one copied from darc, one from svn, and one from git
16:16
<dberkholz>
someone's proposed one before, need to check that out
16:16
<johnny>
so.. i have the quickstart ebuild
16:16
<dberkholz>
i'm sure that git- or darcs-based will be better than svn
16:16
<johnny>
now.. what is the recommended naming for live ebuilds now?
16:16
and then you can joke my poor ebuild coding skillz
16:17
i put it in sys-apps.. since gli was there
16:17
hard to choose between that, app-admin, and app-portage..
16:17
lemme know.. so i know where to commit it to
16:17
<warren>
laga, ah, just asked airlied. apparently upstream currently ati is way ahead in r500 support, r600 is in parity.
16:17
laga, and avivo is pretty much dead
16:18
<johnny>
nice name too :(
16:18
avivo sounds like fun
16:19
ok.. gotta get ready for work.. bbiab
16:24
i'm so ready for spring.. washing dishes is no fun in a sweater :(
16:26mccann has quit IRC
16:28Guaraldo has left #ltsp
16:30
<dberkholz>
johnny: foo-9999.ebuild
16:36elisboa_ has quit IRC
16:45
<laga>
warren: too bad. but two projects are more than enough :)
16:53Egyptian[Home] has quit IRC
16:56Egyptian[Home] has joined #ltsp
17:01gonzaloaf has quit IRC
17:01mhterres has quit IRC
17:01abadger1999 has joined #ltsp
17:09mccann has joined #ltsp
17:11nicoAMG has quit IRC
17:18
<vagrantc>
ok!
17:18
booted NBD+ext2+tmpfs bind mounts
17:18
it *is* faster.
17:18gentgeen__ has quit IRC
17:20
<vagrantc>
now to try with NBD's copy-on-write
17:34
hrm.
17:35
can't seem to actually get it to mount read-write.
17:43ogra_cmpc has quit IRC
17:47abadger1999 has quit IRC
17:58ogra_cmpc has joined #ltsp
18:01johnny_ has joined #ltsp
18:06bobby_C has quit IRC
18:43staffencasa_ has joined #ltsp
18:44Faithful has quit IRC
18:49deavid has quit IRC
18:50
<johnny_>
dberkholz, so when do you think they will have the overlay up?
18:50
<dberkholz>
johnny_: hopefully within a couple days. haven't heard back yet
18:51
<johnny_>
well, i definitely need an ltspfs ebuild for this process to complete ..
18:51
actually, i'm trying to figure out what's calling it..
18:52staffencasa_ has quit IRC
18:54
<ogra>
laga, awake by chance ?
18:55
<warren>
vagrantc, I highly doubt anybody has tested NBD cow lately
18:55
<vagrantc>
warren: yeah, i'm filing a bug
18:55
<warren>
ogra, what is the latest point of a bzr repo called?
18:55
ogra, in in git it is called master
18:55
<ogra>
no idea
18:55
sorry cant discuss atm
18:56
<vagrantc>
warren: could you give more context or more specificity?
18:56
<warren>
just wanting to use the right terminology
18:57
<laga>
ogra: yes.
18:57
<johnny_>
head?
18:57
<warren>
I dunno
18:57
<johnny_>
prolly head..
18:57
at least eveyrbody will know what you're talking about
18:57
no matter which vcs you use
18:57
<ogra>
laga, just wanted to complain that your last patch didnt apply but i noticed it uses different strip values for patch
18:58
hmm, but mythbuntu6 doesnt work ...
18:59
<laga>
the patches don't apply on top of each other
18:59
what's wrong?
18:59
<ogra>
oh
18:59
you should have noted that in the bug
18:59
so you only want -6 applied at all ?
19:00
<laga>
yes
19:00
sorry about the confusion
19:00
i onyl want the latest
19:00staffencasa has quit IRC
19:06
<warren>
vagrantc, ogra: any preferred command for tarball?
19:06
mkdst tar
19:06
?
19:06
mkdst deb
19:06
mkdst rpm
19:06
I guess it works.
19:06
<ogra>
tar is fine
19:09
<vagrantc>
mkdst deb probably should behave similarly to rpm, only checking for a debian dir instead of a spec file
19:09
but i don't forsee using it much
19:09
<warren>
right
19:13
<ogra_cmpc>
i wont use it at all
19:14
<vagrantc>
well ...
19:14
i might play with mkdst deb a bit ... if it can save me doing something manually, i'd be happy for it to do so.
19:15
actually, main thing is renaming the .tar.gz ... NAME_VERSION.orig.tar.gz is the name required for debian packages
19:15
<ogra_cmpc>
well indeed you could fully automate everything :)
19:15
<vagrantc>
rather than NAME-VERSION.tar.gz
19:16
<ogra_cmpc>
well, we should probably just add a -deb option to mkdst tar then that mangles the output filename
19:16
thats something i'd use
19:17
<vagrantc>
i guess .tar.bz2 are accepted now ...
19:17
<ogra_cmpc>
not sure
19:17
none of my packages have .bz
19:17
<vagrantc>
there was an announcement a while back
19:17
at least in debian
19:17
<ogra_cmpc>
we start using lzma all over the palce now
19:18
i think thats still pending debian inclusion
19:18
but itr shrinks openoffice by 15% or so
19:18
<vagrantc>
all those .tar.gz that rarely get downloaded could probably save a good deal of disk space with better compression
19:19
<ogra_cmpc>
disk space is cheap
19:21
<vagrantc>
yes, but spread that across hundreds of mirrors synced daily ...
19:21
cpu for rarely decompressed tarballs is cheaper.
19:22DonSilver has joined #ltsp
19:24
<vagrantc>
so, as far as i can tell NBD copy-on-write is borked.
19:26
<ogra_cmpc>
oh, why ?
19:26
last time i tried it it worked
19:26
<vagrantc>
well, in sid, it appears broken.
19:26
<ogra_cmpc>
ah
19:26
<vagrantc>
1:2.9.9-6
19:26
<ogra_cmpc>
i dont thing we synced it the last month
19:26
*think
19:27
<vagrantc>
actually, it's been there a while ... what version do you have?
19:27
<ogra_cmpc>
1:2.9.9-1ubuntu1
19:28
<vagrantc>
last i tried, the version in etch 2.8.7-4 was totally unsuitable for root fs
19:28
would just hang all the time
19:28
<ogra_cmpc>
i didnt try rootfs
19:28
but for image building it worked
19:29
<vagrantc>
i guess i could try with a different image ...
19:32
<DonSilver>
hi boy & girls (if here) greetings from new york
19:35
<warren>
mkdst tar --release
19:35
Would srccheck and export_tree_vcs
19:35
without --release it only makes a copy of the current contents (even if it isn't checked in)
19:36
<vagrantc>
hmmmmm.....
19:36
with a different image, it appears to work ...
19:37rcc has joined #ltsp
19:41DonSilver has quit IRC
19:43
<vagrantc>
ok ...
19:43
when run from inetd, NBD copy-on-write doesn't work, but when run from the commandline, it seems to work fine.
19:44
when nbd-server is started from the commandline
19:46
<warren>
hmm
19:46
mkdst tar --release
19:46
wouldn't really be a release
19:47captain_magnus has quit IRC
19:47
<warren>
what else to call it
19:48captain_magnus has joined #ltsp
19:48
<johnny_>
snap?
19:48
snapshot?
19:49
<warren>
no
19:49
the default is snapshot
19:49
well
19:49
the default is not from vcs
19:49
so I guess it is snap
19:50rcc is now known as GodFather
19:50
<johnny_>
i didn't actually know what mkdst did, didn't look at it yet ..
19:51ogra_cmpc has quit IRC
19:51
<warren>
Oh, I know.
19:52
it is pointless to implement --snap because it would be the same as the current dir checkout
19:52
i'll instead implement --from-tag=FOO
19:52
so you can reproduce a specific tag tarball exactly
19:53
<vagrantc>
this sounds like pristine-tar
19:53fgiraldeau has joined #ltsp
19:53
<vagrantc>
http://packages.debian.org/pristine-tar
19:58ogra_cmpc has joined #ltsp
20:05patrick_ has joined #ltsp
20:05patrick_ is now known as patrick1345
20:06
<patrick1345>
hello everyone
20:11
is anyone around that could anwser some very basic questions for me about linux and the terminal services
20:13
<jcastro>
just ask
20:16
<patrick1345>
ok I'm kind of a novice at using linux, but I would like to set up a system using 'fat' clients so that I can work on my local machine, but also when I need to log remotly into a remote location so that I can open up large database without it being very slow and using alot of network traffic.
20:16
Right now my company uses exclusivly microsoft servers, but I would like to try to start working on integrating linux.
20:38sbalneav has joined #ltsp
20:39
<sbalneav>
Evening all
20:39
<jcastro>
hi sbalneav!
20:39
<sbalneav>
Hey jcastro
20:39
<ogra_cmpc>
!s
20:39
<ltspbot>
ogra_cmpc: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
20:39
<jcastro>
patrick1345: what I did at my last job was enable a PXE-menu on boot, so that people could reboot the machines, and then "log into" LTSP for things like you're describing.
20:39
<ogra_cmpc>
thats what i call timing
20:40
<vagrantc>
sbalneav: YOU!
20:40
<ogra_cmpc>
2h after ubuntu went into beta freeze
20:40
<jcastro>
patrick1345: the default for the menu was to locally boot the machine (with like a 10 second timer), but if people wanted they could pick "Linux" in the menu and boot the machine in thin client mode
20:40
<ogra_cmpc>
sbalneav, how is life ?
20:40
<sbalneav>
Finally improving
20:41
<ogra_cmpc>
yay
20:41
<sbalneav>
my wife finally got another job
20:41
So, the grieve, and the court battle and everything else is now dropped.
20:41
Hey vagrantc
20:41
<vagrantc>
man, just when i was about to announce happily that the initramfs-tools hooks for nbd-client work almost out of the box with ltsp... sbalneav shows up!
20:41* vagrantc smiles at sbalneav
20:42
<ogra_cmpc>
heh
20:42
<sbalneav>
So, to celebrate, let me post something I've been working on.
20:42
Probably too late for hardy, buuuuuut.
20:42
you can see.
20:42
<ogra_cmpc>
sbalneav, mythbuntu includes dislkes mythtv server/client install based on ltsp now :)
20:42
<fgiraldeau>
sbalneav: Hi there
20:43GodFather has quit IRC
20:43ltsppbot has quit IRC
20:43lilian has quit IRC
20:43dberkholz has quit IRC
20:43
<fgiraldeau>
No, you're kidding!
20:43
<ogra_cmpc>
fgiraldeau, hey, where have you been hiding ?
20:43
<vagrantc>
sbalneav: ldm is feature-complete enough to kill sdm now :)
20:43
<warren>
scotty!?
20:43
<sbalneav>
Yep
20:43
<ogra_cmpc>
suddenly everyone crawls out of their pit ... what happend ?
20:44
<sbalneav>
hold on, lemme push a branch
20:44
gimme 2 secs
20:44dberkholz has joined #ltsp
20:44
<fgiraldeau>
Well, I was very busy... But seems that I'm a survivor
20:44
<warren>
sbalneav, court battle?
20:44
<ogra_cmpc>
yeah seems we all were busy
20:45
<sbalneav>
yeah. Very, very very long story.
20:45* ogra_cmpc tries to remember who recently asked if actually someone in #ltsp works on ltsp :)
20:45
<fgiraldeau>
It's really interesting, we gonna work on NX, memory problems of Firefox and OpenOffice, VoIP for thin-clients, and NBD robustness
20:45
<sbalneav>
My wife got the royal screw job
20:46
<ogra_cmpc>
fgiraldeau, did you try ff 3 beta4 ??
20:46
<fgiraldeau>
sbalneav: What is that
20:46
<warren>
wow, everyone really is here now.
20:46
except Eric
20:46
<ogra_cmpc>
its supposed to solve the memory probs of the world
20:46
<fgiraldeau>
I did try FF 3, but probably not beta4.
20:46
<ogra_cmpc>
that came out last week
20:46
<warren>
sbalneav, I don't know the details, but you have moral support here.
20:47
<fgiraldeau>
I probably should update data about the scenario, to test if the problem is yet there.
20:47* ogra_cmpc offers beer support
20:47
<sbalneav>
What I needed was legal support. But it's over now.
20:47
She started the new job a week and a bit ago.
20:47
so
20:47
To decompress
20:47
<ogra_cmpc>
sbalneav, well, given your joob that shouldnt be hard to find one would think
20:48
<sbalneav>
In winnipeg? IT jobs are hard to come by
20:48
Not a big IT centre here.
20:48
I've branched LDM, and fixed up a TON of things
20:48
<ogra_cmpc>
i meant given *your* job legal help shouldnt be hard to find :)
20:49
<sbalneav>
Phhht. I make too much to qualify for legal aid.
20:49
<fgiraldeau>
sbalneav : So, Why not came in Sherbrooke! ;)
20:49
<ogra_cmpc>
sbalneav, sad that you didnt come yesterday () or three hours ago
20:49
now freeze is in place
20:49
<sbalneav>
That's ok
20:49
Code will be there for next time.
20:49
<fgiraldeau>
Show must go on...
20:49
<sbalneav>
so.
20:50
<ogra_cmpc>
well, i have to support that crap for three yeas, every extra fix would have helped
20:50
<sbalneav>
Well, I've added no new features
20:50
just cleaned up existing ones.
20:50
<ogra_cmpc>
exactly waht i wouldhave needed ...
20:50
but well, over now :) it will work as is
20:50
<fgiraldeau>
ogra_cmpc: There is time yet for bug fixes, right?
20:51
<ogra_cmpc>
fgiraldeau, only beta critical bugs
20:51
<sbalneav>
1) completely glib-ified.
20:51
<ogra_cmpc>
its an lts release
20:51
<fgiraldeau>
yeah
20:51
<ogra_cmpc>
that means we're far more arse about policies
20:51
<sbalneav>
2) 3 different fork/exec methods gone. ldm_spawn now does it all
20:52
<ogra_cmpc>
(and we haved a new release manager who is really blocking if it doesnt apply)
20:52
<sbalneav>
3) major cleanup of ssh_chat, making it (hopefully) language independent.
20:52
and I'm looking right now at fixing the -ac bit.
20:53
the branch is pushing
20:53
<ogra_cmpc>
i'm not sure we can keep the two shh session thing ... i still didnt manage to get CK to work with it
20:53
<fgiraldeau>
ogra_cmpc: that's fair. We will work to support this.
20:53
<ogra_cmpc>
which means no administrative apps at all
20:53
fgiraldeau, the current design is unlikely to work with it
20:54
we extra patches sshd to support CK ...
20:54
i see a session set up ...
20:54
but no way to make it active
20:55
sshd approves the first connection, not the second
20:55
<sbalneav>
vagrantc: so, your guest login bit doesn't actually pop up a button? Just a blank entry in the username field?
20:55GodFather has joined #ltsp
20:55lilian has joined #ltsp
20:55
<warren>
what is CK?
20:55
<ogra_cmpc>
consoleki6t
20:56
<warren>
oh, I see.
20:56
sbalneav, will your fixes go into ldm-trunk?
20:56
<ogra_cmpc>
the stuff that IMHO has gone into gnome 1year to early
20:56
<sbalneav>
ogra_cmpc: So, for CK, you want only one ssh login?
20:56patrick1345 has quit IRC
20:56otavio has quit IRC
20:56
<warren>
ogra_cmpc, I'm not too happy with it either
20:56
<ogra_cmpc>
sbalneav, well, not sure that fixes it yet
20:57
i need to do more debugging
20:57
<warren>
sbalneav, where are your ldm fixes now?
20:57
<sbalneav>
warren: Well, that's up to everyone else! I'll post a branch, and people can look and see. I don't want to move it to -trunk until people think it's worthwhile.
20:57
warren: uploading.... :)
20:57
<ogra_cmpc>
but given that i only had about three four days at all this release for ltsp its in pretty bad shape
20:57
<warren>
sbalneav, cool ok
20:57
<sbalneav>
bzr push sftp://bazaar.launchpad.net/~sbalneav/ltsp/ldm-scott
20:58
<warren>
let us know when it is done
20:58
<sbalneav>
\ | / - \ | / -...
20:58
<ogra_cmpc>
its gutsy with as many fixs as i could pull in
20:58
<warren>
too bad we didn't start the ldm repo fresh, it would be a lot smaller/faster now
20:58
<sbalneav>
(that was the bzr spinner)
20:58
<ogra_cmpc>
ugh
20:58
dont use sftp
20:59
use bzr+ssh://
20:59
<sbalneav>
oh, ok!
20:59
<ogra_cmpc>
thats nearly twice as fast
20:59
<sbalneav>
I've always used the sftp :)
20:59
<ogra_cmpc>
i think zou can also just use lp://
21:00
<sbalneav>
warren: saw your last bit about -ac, and you're right on the money
21:00
<ogra_cmpc>
that should translate to -> bzr+ssh//bazaar.launchpad.net/
21:00
<warren>
sbalneav, I was right, but I used his patch verbatim =)
21:00
<ogra_cmpc>
sbalneav, we spent some nights with it here
21:01
<sbalneav>
Oh, is it fixed somewhere else other than in trunk, or still not fixed yet?
21:02ltsppbot has joined #ltsp
21:03
<ogra_cmpc>
sbalneav, that needs deeper thinking about proper xauth usage ...
21:03
<sbalneav>
right.
21:03
<ogra_cmpc>
i thought vagrant submitted the -ac removal though
21:03
<sbalneav>
yeah, he did.
21:04
I was going to look at doing an ldm_login script to pass the terminal's xauth key up to the server.
21:04
do it properly.
21:05
still chugging.
21:05
<ogra_cmpc>
you mean an rc.d script ?
21:05
<sbalneav>
right.
21:05
<ogra_cmpc>
yeah sounds good
21:05
<sbalneav>
something like:
21:06
XAUTHKEY=$(xauth --grab_my_key)
21:06
<ogra_cmpc>
looks good
21:06
<sbalneav>
ssh -S $LDM_SOCKET $LDM_HOST xauth add $LDM_IPADDR:$LDM_VTY-1 $XAUTHKEY
21:06
or the like
21:07
along those lines.
21:07
vagrantc: still there?
21:07
<ogra_cmpc>
sbalneav, http://people.ubuntu.com/~ogra/classmate/images/hardy/
21:07
something to test for you btw
21:10
<sbalneav>
K, I'll snag it
21:11
<ogra_cmpc>
its quite usable :)
21:13spectra has quit IRC
21:15
<fgiraldeau>
I see the situation. I will get into the new code and give you feedback on mailing list saturday.
21:15
Take care all, I'm going to sleep
21:15fgiraldeau has quit IRC
21:16DonSilver has joined #ltsp
21:16
<sbalneav>
https://code.edge.launchpad.net/~sbalneav/ltsp/ldm-scott
21:17
<ogra_cmpc>
sbalneav, edge isnt good outside of ubuntu ... for
21:17
public usage you should rather post with the edge ripped out
21:18
(it requires certain team privileges iirc)
21:20
wow
21:20
quite a diff
21:20
<warren>
do you prefer that size of patch or my billion patches? =)
21:21
nevermind
21:21
<sbalneav>
Lots of it is the glibification
21:21
(is that a word?)
21:21
glibossitude
21:22
<warren>
sbalneav, that's good
21:22
<sbalneav>
glib-inally, it's much better
21:22
<warren>
sbalneav, do you know what's the minimum glib API you coded to?
21:22
<ogra_cmpc>
warren, its called "major rewrite" :)
21:22
<sbalneav>
ummm
21:23
The fanciest thing I'm doing is g_spawn_with_pipes
21:23
so, it shouldn't require latest-and-greatest, I wouldn't think.
21:23
the rest is all using g_string functions
21:23
<ogra_cmpc>
warren, worst case the one thats in gutsy
21:24
which is half a year old
21:24
<sbalneav>
And there's a joke in there which, due to the family nature of this channel, propriety dicates I won't tell.
21:24
<warren>
I just pushed a redesign of mkdst
21:24
Unless I find more bugs this should be mkdst-0.6
21:24
I'm testing it on ltspfs now
21:24hari has joined #ltsp
21:24
<warren>
anyone else plan on pushing changes for ltspfs today?
21:25
using ltspfs because it is low traffic
21:27GodFather has quit IRC
21:28
<ogra_cmpc>
warren, not me
21:28
<vagrantc>
warren: not i
21:29
sbalneav: sorry, was away
21:29* vagrantc reads backlog
21:29
<sbalneav>
vagrantc: I can relate :)
21:29* ogra_cmpc needs sleep ....
21:29
<sbalneav>
Night!
21:29
<ogra_cmpc>
night, good to have you back :)
21:29
<sbalneav>
Nice to be back
21:29* ogra_cmpc hugs scottie
21:29
<hari>
i'm using ltsp in edubuntu 7.04 feisty. what billing application is usually used in edubuntu feisty? regards
21:29* sbalneav give ogra_cmpc a goodnight kiss :)
21:30
<ogra_cmpc>
:)
21:30
<sbalneav>
vagrantc: So, guest login
21:30
blank userid?
21:30
I've *&%*$ed that up in my branch.
21:32
<warren>
oh, found a bug in my checkin
21:35
<vagrantc>
sbalneav: RyanRyan52 did a lot of the LDM coding that i committed
21:35
sbalneav: but it does pop up a button ... doesn't actually insert a username ... basically just sets it to behave like autologin
21:35
sbalneav: and autologin now has some "reasonable defaults"
21:37
<sbalneav>
yeah, saw those.
21:37
OK, I've screwed up the guest login in my branch.
21:37
I'll have to sort it out.
21:37
<vagrantc>
i did switch it to LDM_GUESTLOGIN ... was LDM_ALLOW_GUEST for a while
21:38
<sbalneav>
I'll fiddle with it.
21:38
and make it work with the re-worked ldm
21:38
what I'm REALLY interested in seeing people test
21:38
is the new ssh chat logic
21:39
I *HATED* the "talk to ssh in LANG=C and transliterate to local lang" bit.
21:39
it was sucky
21:39
and hacky
21:39
and icky
21:39
<vagrantc>
ssh is localized?
21:39
<sbalneav>
I believe so.
21:39
the
21:39
the "passwd" program it calls to change passwords certainly is
21:39
<vagrantc>
ah
21:40
<sbalneav>
so now, what I do is simply search for either ": " (i.e. a password prompt)
21:40
or LTSPROCKS sentinel
21:40
<warren>
what if LTSPROCKS is my password?
21:40
<sbalneav>
and I pass through the messages that ssh spits out to the message area
21:40
<vagrantc>
if the server is echo'ing your password, it's bad news.
21:40
<warren>
=)
21:40
<sbalneav>
so... *IN THEORY*,
21:41
warren: no, it doesn't :)
21:41
if you've set up your language to lower slobovian, and ssh into the server, and you have a password expiry
21:41
ssh will talk in slabovian, and pass that through to the greeter.
21:41
<vagrantc>
it took me quite a while to realize that LTSPROCKS wasn't just an easter egg
21:42
<sbalneav>
the only thing I rely on is that password prompts end with ": "
21:42
which is better than it is now.
21:42
but we need non-english speakers to test that.
21:42
<vagrantc>
is it possible that anything else could end in ": " ??
21:42
<warren>
en_CA isn't non-english?
21:43
<sbalneav>
Hush, American :)
21:43
<vagrantc>
would hate to send the password without being absolutely sure it was actually a password prompt
21:43
<sbalneav>
the ": " has to be the last thing on the line.
21:43
<warren>
it runs passwd?
21:43
vagrantc, no way to guarantee it by screen scraping alone, there are multiple different implementations of passwd
21:43
<sbalneav>
sshd runs passwd if there's been a pam_ password expired.
21:44
it's actually hard coded into sshd
21:44
<warren>
oh, this passwd is?
21:44
<sbalneav>
yeah
21:44
if you've got PAM support enabled in ssh, that's what it runs to handle an expired password.
21:45
<warren>
does it actually connect with a console, or is this implemented by the client talking a the protocol?
21:45
(client itself gives the prompt?)
21:45
<sbalneav>
actually connects with a console.
21:45
hence the reason why I have to screen scrape ssh
21:46
if there was a *%^$*&^ api to do it, I'd LOVE it.
21:46
<warren>
hm..
21:46
I suppose this is OK for a first implementation
21:46
<sbalneav>
what you'd REALLY like to have happen is use the ssh-askpass mechanism for expired passwords as well.
21:46
<warren>
please mark it with /* XXX: HAAAAAAAACK! Redo this without screen scraping with proper l10n */
21:47
<vagrantc>
sbalneav: so have you secretly been preparing this glibification branch, or did you just whip it together in the last few minutes or something? :)
21:47
it's quite the diff
21:47
<sbalneav>
I've done this all in the last 4 days.
21:47
<warren>
sbalneav, when we later l10nize ldm with gettext, we can more reliably scrape LANG=C and print a l10n prompt, unless we find a better way.
21:48
oh crap, that might make it difficult to login with a non-C lang?
21:49
<sbalneav>
right, which is why I'm now passing through the prompts.
21:49
<vagrantc>
sbalneav: i can upload ltsp, ldm and ltspfs directly now ... so i could do an upload to debian experimental and get some testers
21:49
<warren>
sbalneav, I can test this in fedora as well
21:50
<vagrantc>
and virtualbox allows me to get work done even in total isolation :)
21:50
<warren>
so xauth isn't fixed yet?
21:51
<sbalneav>
No, I'll have a look at xauth this weekend.
21:51
<warren>
sbalneav, ok cool
21:51
<sbalneav>
it's a big problem and should be fixed properly.
21:51
<vagrantc>
sbalneav: one of the issues with the xauth thing is that using ssh -X connects to an X proxy, and actually having the authorization key would be rejected by the proxyt
21:51
<warren>
sbalneav, I might be checking in changes to support mkdst-0.6 in ldm tonight
21:51
<sbalneav>
ok
21:51
<warren>
sbalneav, actually the only change might be to delete ldm.spec
21:51
i'll see soon
21:52
<vagrantc>
i've struggled with this xauth thing ... no clue where it's failing, really.
21:53
sbalneav: why TRUE changed to 1 and FALSE changed to 0 ?
21:54
<sbalneav>
Laziness. I need to fix that :)
21:54
But you'll notice I hacked out all the #includes in that file :)
21:55
<vagrantc>
yes. i was thinking we could just apply that patch.
21:55
<warren>
I'll do a quick test of his patch right now
21:56
<vagrantc>
i'd like to try and merge some of the changes piece by piece ... a lot of it is just glib stuff, but some of it could be isolated out
21:56
i can work on that tomorrow, most likely.
21:58
<warren>
grr, don't have a chroot installed
21:59
will take a few minutes to install
22:03
<vagrantc>
sbalneav: looks like you removed the if(ldminfo.autologin); load_guestinfo(); call ...
22:04
<sbalneav>
Yeah, like I said, I mucked that up.
22:04* vagrantc is just trying to figure out where
22:04
<sbalneav>
I had inlined some of it, and it needs to go back out to a separate function.
22:04
like you had it.
22:05
I had started all of this with a gutsy tree, then this morning tried merging in the latest ldm-trunk stuff.
22:05
<vagrantc>
whoah. :)
22:05
<sbalneav>
Don't worry about it tonight
22:05
I'll fix it up tomorrow morning.
22:06
<vagrantc>
overall, looks a lot cleaner
22:06
<sbalneav>
Figured if I was going to show up, I didn't want to come empty handed :)
22:06
<vagrantc>
:)
22:06
<sbalneav>
yeah, it's a definite improvement.
22:06
not so much "i++" now :)
22:06
<vagrantc>
sbalneav: with you away, i've actually learned a bit of C ... working on ldm and ltspfs
22:06
<sbalneav>
good.
22:07
I've learned a lot about union greivance processes.
22:07
<vagrantc>
ouch.
22:07
<sbalneav>
I'll wager which one's the more useful knowledge :)
22:07
hint
22:07
<vagrantc>
depends on your situation, most likely :)
22:07
<sbalneav>
not the grievence stuff
22:08
anyway
22:08
alice's 3 year long ordeal's finally over.
22:08
<vagrantc>
6 files changed, 529 insertions(+), 747 deletions(-)
22:09
over 100 lines of code gone! :)
22:10
<sbalneav>
smaller = better :)
22:10
KISS
22:10
err
22:10
smaller == better
22:10
Don't want any side effects in an if statement :)
22:11
<vagrantc>
heh
22:12* warren is too curious about what happened with that legal thing.
22:12
<warren>
sbalneav, ok, well, I built with your patch and it seems to be working with fedora 8
22:13
<sbalneav>
meh, too long to type :)
22:13
work being a jerk to my wife
22:13
<long boring greivance/court process>
22:13
<warren>
oh
22:13
<sbalneav>
<work realizes they've screwed up>
22:14
Wife gets moved to different department with old job responsibilities intact
22:14
as of 1.5 weeks ago.
22:15
<warren>
and you're stuck with the 3 years of legal bills
22:15
?
22:15DonSilver has quit IRC
22:15
<sbalneav>
it was the old "we'd like to eliminate a position, so to save ourselves the package we'd have to pay her to lay her off properly, we'll cook up a case that she's incompetent so that we can just fire her"
22:16
no, no bills, union lawyer was handling it.
22:16
blah blah blah tl;dr
22:17
glad it's over is all I gotta say.
22:17
<warren>
so instead they pay 5x as much in legal expenses trying to let her go?
22:18
<sbalneav>
sure
22:18
Working for government's kinda odd sometimes.
22:18
My wife works for the city, I work for the province.
22:20
Anywho. Tired. I'll be on tomorrow.
22:20
I'll fix up the autologin stuff, vagrantc
22:20
night all.
22:20sbalneav has quit IRC
22:21* warren too
22:25johnny_ has quit IRC
22:32mccann has quit IRC
22:35hari has left #ltsp
22:46MacIver has quit IRC
22:53lns has quit IRC
23:20
<warren>
vagrantc, hm, I might not be able to USE start_sound in the common script because newer pulseaudio needs different parameters to launch
23:22MacIver has joined #ltsp
23:27ogra has quit IRC
23:27ogra has joined #ltsp
23:32
<vagrantc>
warren: i'm thinking start_pulseaudio can be defined in the common script, and then we have the possibility of overwriting it in our own scripts ... ideally working towards merging.
23:32
<warren>
vagrantc, overrding you mean?
23:32
<vagrantc>
yeah, overriding
23:33
<warren>
so source it first
23:33
then write your own
23:33
and it'll use the second one
23:33
<vagrantc>
exactly.
23:33
<warren>
ah, simple. k.
23:34
<vagrantc>
wondering if we should ever deprecate esd and nas ...
23:35
<warren>
well, consider who is using LTSP5 in trunk now
23:35
<vagrantc>
well, ubuntu is using pulse, debian is using pulse ... fedora's going to be using pulse ?
23:36
although it's a tiny enough bit of code that's not going to change much ... not so evil to leave it in
23:36
<warren>
fedora has been using pulse by default in NORMAL desktop sessions even intercepting ALSA since Fedora 8
23:37
leave it for now
23:37
our spyware hidden in ltsp-server can tell us how many people are using it
23:37
=)
23:37
<vagrantc>
heh.
23:38
debian has opt-in spyware: http://popcon.debian.org
23:38
<warren>
I really like this new mkdst
23:38
<vagrantc>
331 users of ltsp, 73 of ldm, 163 of ltspfs ...
23:39
<warren>
I can continue doing what i've been doing before, except my spec changes go elsewhere
23:39
<vagrantc>
good.
23:39
the one down-side is ... sometimes i'd actually see something in your .spec file and realize i should do the same for debian
23:40
<warren>
well, we still need to do more tree organization changes
23:40
<vagrantc>
warren: but i'm guessing you'll publish your spec file in a way that i can easily keep track of it?
23:40
<warren>
but I guess getting the code in first makes the tree changes more obvious
23:40
vagrantc, like hell
23:40* vagrantc raises an eyebrow
23:41
<warren>
hmm
23:41
vagrantc, I'm super tired and the stupid humor is really kicking in
23:41* warren trying to find URL
23:41
<vagrantc>
well, i'd best get some sleep myself.
23:41
<warren>
vagrantc, http://cvs.fedoraproject.org/viewcvs/devel/ltsp/ I think that's a permalink
23:42
<vagrantc>
warren: cool.
23:42
<warren>
vagrantc, oh, let me know if you have any ideas how we can implement each mkdst command as its own file
23:42
vagrantc, so we can make it into plugins instead of hard coded into mkdst
23:42
although maybe doing that would be a bad idea
23:42
i dunno
23:43
because then you would need to create a stable API
23:43
if you can have a bash API
23:43
<vagrantc>
i like plugin type things, so i'll definitely weigh in on it
23:43
tomorrow :)
23:43
<warren>
good night
23:43vagrantc has quit IRC
23:49cyberorg has quit IRC
23:56rafi has joined #ltsp