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


Channel log from 22 January 2010   (all times are UTC)

00:06kaos01 has quit IRC
00:07alkisg has quit IRC
00:12alkisg has joined #ltsp
00:20
<alkisg>
vagrantc: wouldn't it be better for ltsp-build-client.conf if $EARLY_MIRROR and $EXTRA_MIRROR were enter-separated instead of comma-separated? (of course not the equivalent command line options, those should stay comma separated...)
00:21
Good morning all, btw :)
00:23
E.g. now I need this in ltsp-build-client.conf: http://paste.ubuntu.com/360478/
00:25
Or, could I just commit something that would make both separators acceptable?
00:26
(hmmm yup that sounds like a better idea :))
00:29
Heh, it's even smaller: http://ltsp.pastebin.com/m12765f0
00:35
Test case: http://ltsp.pastebin.com/m1a60832a - seems good enough, commiting...
00:37
Uhm /me looks at why Ubuntu has a different file there than Debian... :D
00:43
hmok, committing the change for both distros, I hope that's the right thing to do.
01:00
<vagrantc>
alkisg: i think enter is just whitespace in variables...
01:01
<alkisg>
vagrantc: if you do `echo "$variable"`, in quotes, it stays intact
01:01
if you do `echo $variable` without quotes, all spaces/enters are consolidated
01:02
I tested it with sh+bash before committing, it seems to work fine...
01:02
<vagrantc>
ltsp-build-client is all bash anyways
01:20
<alkisg>
(so btw in the old code, OIFS=$IFS would "destroy" $IFS as it wouldn't keep the enters/tabs there - OIFS="$IFS" would work better)
01:42prpplague_afk has quit IRC
01:43
<vagrantc>
sure.
01:47
alkisg: that looks much simpler :)
01:47
<alkisg>
Thanks :)
01:50
<vagrantc>
we could probably sync more of the code between those two functions rather than fork it...
01:51
<alkisg>
Yes, I don't think two different 010-manage-mirrors are _needed_, it'd be easier to maintain them in a single file
01:55
<vagrantc>
just need to figure out a clean way to handle variances.
01:55
<alkisg>
Also, I don't think there's a need for 2 different functions (add_multiple_mirrors / add_mirror), one add_mirrors function would be enough
01:55
(always accepting one variable, either comma or enter seperated if it has many mirrors)
01:56
<vagrantc>
i forget why i had two different functions... it made sense at the time.
01:56
might have been as simple as keeping the indentation simpler
01:56prpplague_afk has joined #ltsp
01:56
<alkisg>
| while read mirror; do ==> would just become ==> | while read mirror type components; do
01:57
awk would also not be needed this way
01:58Selveste1___ has quit IRC
01:58* alkisg tries this...
02:02* vagrantc sleeps
02:02vagrantc has quit IRC
02:02
<alkisg>
vagrantc: http://ltsp.pastebin.com/m7a9e6ca2
02:02
Heh, too late :)
02:05gnunux has joined #ltsp
02:09tarbo has joined #ltsp
02:15
<gnunux>
hi
02:36grey-monkey has quit IRC
02:42Egyptian[Home] has quit IRC
02:45Egyptian[Home] has joined #ltsp
02:55bobby_C has joined #ltsp
03:05evilx has quit IRC
03:25rhodan_ has quit IRC
03:54sene has joined #ltsp
03:59Selveste1 has joined #ltsp
04:24rhodan has joined #ltsp
04:36mikkel has joined #ltsp
04:44alexqwesa_ has joined #ltsp
04:57alexqwesa has quit IRC
05:01hersonls has joined #ltsp
05:20otavio has joined #ltsp
05:42Silveira_Neto has joined #ltsp
05:42
<Silveira_Neto>
Hi guys. I'm trying to enter in the main ltsp user list, how I do that?
05:49
<alkisg>
Silveira_Neto: what do you mean? LTSP uses whatever users you have defined in your server...
05:49
<Silveira_Neto>
I say, LTSP email user list. :=)
05:50
<alkisg>
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
05:51
<Silveira_Neto>
Thanks!
05:51
<alkisg>
You're welcome
05:52
!learn ltsp-discuss as at https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
05:52
<ltspbot>
alkisg: The operation succeeded.
05:52
<alkisg>
!learn ltsp-developer as at https://lists.sourceforge.net/lists/listinfo/ltsp-developer
05:52
<ltspbot>
alkisg: The operation succeeded.
05:52
<alkisg>
!ltsp-developer
05:52
<ltspbot>
alkisg: "ltsp-developer" :: at https://lists.sourceforge.net/lists/listinfo/ltsp-developer
05:58ogra has quit IRC
05:58ogra has joined #ltsp
06:02Silveira_Neto has left #ltsp
06:24pmatulis has joined #ltsp
06:26scottmaccal has joined #ltsp
06:52dro has joined #ltsp
06:52
<dro>
pmatulis: good morning, good news.....I tested the isync setup and it worked
06:53
oops *iscsi
06:53
<pmatulis>
dro: super, i'm sure you found something to change on the page?
06:53
dro: something to improve, add info, etc
06:53
<dro>
pmatulis: yes, nothing major
06:54
pmatulis: my test server is extremly crappy, so I'm going to rebuild both my vm's tonight
06:54
and then document a little more heavily on the setup
06:55
for example: how to add a 2nd disk, how to tell what your current disks are /dev/sdc /dev/sdb
06:55
<pmatulis>
dro: ok
06:55
<dro>
pmatulis: but overall it's badass
06:55
<pmatulis>
dro: heh
06:55
<dro>
pmatulis: you have no idea how much i appreciate it
06:56
<pmatulis>
dro: well let's make it better!
06:56
<dro>
sounds good to me
06:56
<pmatulis>
dro: did you also need to use rc.local to mount ecryptfs? or did you not try?
06:57
to use fstab
06:57
<dro>
pmatulis: I didn't try to use fstab
06:57
i just got it working and that was it
06:58F-GT has quit IRC
06:58
<pmatulis>
dro: ok
06:58
<dro>
pmatulis: I know you wanted to know about that but the main reason that I didn't is I wasn't sure how to add it to fstab
06:58
I did the quick and dirty run through install
06:59
the 2nd time around I'm actually going to research the options and read the man pages
06:59
<pmatulis>
dro: alright
06:59
dro: i'm looking forward to seeing your changes on the page
07:00Selveste1 has quit IRC
07:00
<dro>
pmatulis: cool they will be there this weekend!
07:01
ok i'm off to work
07:01dro has quit IRC
07:05F-GT has joined #ltsp
07:09rhodan_ has joined #ltsp
07:25rhodan has quit IRC
07:30alkisg has quit IRC
07:32bobby_C has quit IRC
07:48litlebuda has joined #ltsp
07:59lukas2511 has joined #ltsp
08:01
<lukas2511>
hiho.. can someone please help me.. i have a problem with ltsp5.. all sessions are very slow and i think that the problem is the ssh connection to the clients cause ssh is a little bit slow on my server....
08:02
<gnunux>
why ssh is "little bit slow" on your server ?
08:03
<lukas2511>
i dont know.. if i open up to 5 ssh sessions everything is ok but if i open more than 5 the reaction gets a little bit slow.. tested with "htop" in all sessions at the same time over a gigabit network connection
08:04
the cpu is a AMD Phenom II X4 920 and the machine runs with 8gb of ram
08:04
it's a "clear" debian installation with ltsp-server-standalone and ssh
08:12rhodan_ has quit IRC
08:13
<lukas2511>
ouh.. other admin wants to install another distro so.. maybe see ya later xD
08:13lukas2511 has quit IRC
08:15cliebow has joined #ltsp
08:16etyack has joined #ltsp
08:30prpplague_afk is now known as prpplague
08:45alkisg has joined #ltsp
08:46dro has joined #ltsp
08:46
<dro>
good morning, again lol
08:52
<sbalneav>
Morning all
08:53
<dro>
sbalneav: good morning
08:58
<alincoln>
morning
09:01coordinador has joined #ltsp
09:01
<coordinador>
hi
09:02
i have lts1200 term, but i want to know how to use the right video drivers
09:02
because, despite im using firefox and flash locally, still works slowly
09:03
<alkisg>
coordinador: what video card do those have?
09:04
<coordinador>
i dont know, how can i know? these are the terminals that sells disklessworkstations
09:04
<alkisg>
What distro/version do you have?
09:04
<coordinador>
ubuntu 9.04
09:04
<sbalneav>
coordinador: on a shell on the thin client:
09:04
lspci && free
09:05
pastebot the results
09:05
<alkisg>
coordinador: on a *local* shell...
09:05
I.e. one you can get with `ltsp-localapps xterm`, or by putting SCREEN_02=shell in lts.conf...
09:06
<coordinador>
alkisg, yes, i did localapps thing :) thanks
09:06
I get: VGA compatible controller: VIA Technologies, Inc. CN700/P4M800 Pro/P4M800 CE/VN800 [S3 UniChrome Pro] (rev 01
09:07
<alkisg>
coordinador: now run lspci -vv -nn, you'll see what driver it uses
09:08
<sbalneav>
I'd suggest making sure you've got the openchrome driver installed in the chroot
09:08
xserver-xorg-video-openchrome
09:08Pod2 has joined #ltsp
09:11CAN-o-SPAM has joined #ltsp
09:13litlebuda has quit IRC
09:16
<coordinador>
oh, ill check that
09:17
when i do "lspci -vv -nn i get" "00:00.0 Host bridge [0600]: VIA Technologies, Inc. CN700/VN800/P4M800CE/Pro Host Bridge [1106:0314]" and other things
09:18
<alkisg>
Don't you see something like the following?
09:19
Kernel driver in use: nvidia
09:19
Kernel modules: nvidia-current, nvidiafb
09:19
<coordinador>
i will grep "kernel driver"
09:19
i get many lines, i cannot paste here
09:19
<alkisg>
4-5 lines after the: VGA compatible controller: VIA Technologies, Inc. CN700/P4M800 Pro/P4M800 CE/VN800 [S3 UniChrome Pro] (rev 01
09:20
<coordinador>
oh, ill check that
09:21
i ll paste in pastebin
09:24
http://pastebin.com/d19dd0ebb
09:25
and with the lspci... command i get: http://pastebin.com/d39f83f68
09:26
sbalneav, and yes, i have installed xserver-xorg-video-openchrome
09:27
<alkisg>
coordinador: so e.g. if you do `ls -R /` in that local terminal, you see the screen scrolling slowly?
09:27
E.g. 5 fps?
09:28
<coordinador>
no, it runs fast
09:28
<alkisg>
Then it wouldn't be a driver issue
09:28
How much ram do you have on the client?
09:29
And what CPU does it have?
09:29
<coordinador>
they have 1gb of ram and 1ghz
09:30
<alkisg>
Yeah that should work fine...
09:30
<sbalneav>
Well, I guess my question is when you say the run slowly
09:30
define "slowly"
09:30
Here's a test
09:31
open up Totem:
09:31
<coordinador>
slowly, means that flash like youtube runs choppy
09:31
<sbalneav>
Aplication->Sound and Video->Movie Player
09:31
<coordinador>
remotely and locally
09:31
<alkisg>
youtube mainly uses the CPU
09:31
<coordinador>
ok sbalneav
09:32
ok
09:32
alkisg, i though that 1ghz cpu could run fine a youtube vid
09:33
<alkisg>
!flash
09:33
<ltspbot>
alkisg: "flash" :: Yes, flash sucks. Make sure you have LDM_DIRECTX=True in your lts.conf file, or if it's just youtube you're after, try the HQtube plugin. Install greasemonkey for firefox, and see http://userscripts.org/scripts/show/24999
09:33
<alkisg>
...and we should add the HTML5 youtube site there probably... :)
09:35
coordinador: well, try "top" locally while flash is running to verify if this is a CPU problem or not.
09:36
<coordinador>
ok, alkisg, ill do that
09:36
<Appiah>
hehe
09:37
I just found some hours ago that the HTML5 youtube existed and been using it
09:37
thinking "Must set this to default on LTSP"
09:39
<sbalneav>
On the right hand side bar, drop down the dropdown on totem, and select youtube
09:39
search for a video (if you search for "Balneaves" you'll get me freeing a bird in spain)
09:39
see how that works.
09:39
<Appiah>
worked bad last time I tried
09:40
and people wanted to use youtube from the browser like they always do
09:40
<alkisg>
There's also a Qt app that acts as a youtube front end
09:40
<coordinador>
i got an error
09:40
<alkisg>
Much more attractive than totem
09:40
<coordinador>
maybe you dont have permission to open this file
09:41
is the message that i get
09:41
<sbalneav>
in what, totem?
09:42
<coordinador>
yes, i get only errors
09:42
<sbalneav>
what errors, exactly
09:42
<coordinador>
"Gstreamer found and error in a general support library"
09:42Selveste1 has joined #ltsp
09:42
<coordinador>
and "maybe you dont have permission to open this file" (with the sbalneav vid)
09:42
<sbalneav>
You've got something badly mucked up with your gstreamer setup then
09:43
are all your packages up-to-date?
09:43
<coordinador>
:(, ok ill do the top command while localapp youtube playing
09:45
OMG runs so slow
09:46
in top i get 90% cpu used by firefox
09:46
<alkisg>
how much idle?
09:46
<sbalneav>
coordinador: So, there's your problem. Flash sucks.
09:46
<coordinador>
"0,0%id" this means idle?
09:46
:(
09:46
<sbalneav>
yeah, 0% idle
09:46
<alkisg>
Yup, that means your CPU is your problem
09:47
<sbalneav>
coordinador: I have a dual core box here:
09:47
model name : Intel(R) Core(TM)2 Duo CPU E8500 @ 3.16GHz
09:47
bogomips : 6344.98
09:47
bogomips : 6344.98
09:47
and flash is slow.
09:47
<coordinador>
:S
09:47
<sbalneav>
and it's a full workstation.
09:47
flash. just. sucks.
09:48
fix your gstreamer stack and get either totem or the QT app going to watch youtube videos.
09:49
I watch youtube videos via totem on thin clients, fullscreen, NOT localapps
09:49
smooth as silk
09:49
<coordinador>
:o
09:49
<sbalneav>
no problems at all, only about 5% cpu load.
09:50
Play 1 youtube video using flash, 100% cpu.
09:50Faithful has quit IRC
09:50
<sbalneav>
can't do fullscreen, can't get it to not be choppy.
09:51Faithful has joined #ltsp
09:51
<coordinador>
then the solution could be the hqtube right?
09:52
<sbalneav>
well, the only problem with hqtube is, you have to install it for EVERY user
09:52
since greasemonkey doesn't support a common script area.
09:59
<coordinador>
humm
09:59
ok
09:59
ill try
09:59
see you
09:59
thank you i have to fly
09:59
byeee
09:59coordinador has quit IRC
10:21tstafford has quit IRC
10:23
<sbalneav>
Yet another disatisfied flash customer :(
10:24* sbalneav shakes fist at adobe in impotent rage
10:24* CAN-o-SPAM adds a fist
10:25* ogra adds HTML5
10:27gnunux has quit IRC
10:27
<sbalneav>
yeah, that'll be great once it gets widespread
10:27tstafford has joined #ltsp
10:28
<sbalneav>
I really can't understand how you can watch, like, 20 youtube videos on the SERVER via totem, and have it work fine, but ONE freaking instance of flash watching a video pegs the cpu.
10:28
I mean, you gotta WORK at it to get an application that inefficient!
10:29
<CAN-o-SPAM>
right?
10:29sene has quit IRC
10:29
<CAN-o-SPAM>
and then lets talk about adobe air ... sheesh
10:29
<sbalneav>
What the heck are they doing? Calculating Pi to 20,000 decimal places in between each frame?
10:29
geez
10:29
<CAN-o-SPAM>
Adobe ... "Inflating the hardware market since 1982"
10:30
<sbalneav>
They're also the one f***ing up printing for me with cups.
10:30F-GT has quit IRC
10:31
<sbalneav>
http://www.cups.org/str.php?L3417+Qversion:1.4
10:31F-GT has joined #ltsp
10:31
<sbalneav>
"Yeah, even though you've fixed the bug, we don't want to apply your patch on the off chance that it may affect adobe apps on Mac. Tough cookies"
10:33
<CAN-o-SPAM>
sbalneav: want to start a new group? LUAAD - Linux Users Against Adobe Development
10:33
<sbalneav>
So, basically, OpenOffice.org printing is broken for EVERYONE on EVERY version of Linux, and we won't fix it because we don't want to possibly affect some Madison Avenue copyrighter producing ads in illustrator.
10:35staffencasa has joined #ltsp
10:35
<sbalneav>
I'm usually a happy drunk, but since that comment's been posted, there's a couple of times when I've had a few beers in me that I've though of forking freaking cups, with the emphasis on making sure it works right for LINUX users, as opposed to apple users.
10:35
>< this close
10:35
what's that, 1 pixel? :)
10:35
2, maybe?
10:35
<alincoln>
hah
10:49
<dro>
lol
10:58
fun stuff....I found an awesome donation of 60 dell optiplex sx280's. we already have 100+ deployed using ltsp and they all work fine. this must be a slightly newer revision because the sound doesn't work! which is the same problem i had originally with the sx 280's that I already have
11:04
<johnny>
sbalneav, yeah.. buying cups.. that sucked..
11:04
have they atually added much of value since they aquired it
11:05
did the web gui ever get rewritten?
11:07
sbalneav, it seems that you should gather reps from every major distro (and invite anyone else of course) if you ever decide to try it. see if it will ever get accepted upstream and whatnot
11:11pmatulis has quit IRC
11:12pmatulis has joined #ltsp
11:20rhodan has joined #ltsp
11:22
<dro>
sbalneav: what would you name your fork? sups?
11:28_UsUrPeR_ has quit IRC
11:35rhodan has quit IRC
11:35rhodan has joined #ltsp
11:37_UsUrPeR_ has joined #ltsp
11:46rhodan has quit IRC
12:00Pod2 has left #ltsp
12:03vagrantc has joined #ltsp
12:05korcan has quit IRC
12:07
<vagrantc>
alkisg: i don't think your proposed add_mirrors will work for debian-style security mirrors.
12:07
<alkisg>
vagrantc: why not?
12:08rhodan has joined #ltsp
12:08
<vagrantc>
alkisg: it would require manually specifying the dist portion, as the dist is $DIST/security ... security is not a component.
12:08
<alkisg>
vagrantc: add_mirrors "url $DIST/security"
12:09
I don't understand where the problem is, can you paste the whole security line?
12:09rhodan has quit IRC
12:09
<vagrantc>
alkisg: deb http://security.us.debian.org/ lenny/updates main contrib non-free
12:10
alkisg: so the old code could specify http://security.us.debian.org
12:10
alkisg: and it would figure the rest out.
12:10* alkisg looks...
12:10
<alkisg>
btw, what's that "none" for?
12:10
I couldn't understand it so I just left it there...
12:10
<vagrantc>
alkisg: some dists don't have security mirrors
12:11
alkisg: overall, i think your approach is cleaner, we'll just need to figure out some special casing for debian's security mirror handling, i think.
12:11
<alkisg>
vagrantc: ah, I got it, you pass "security" as a flag
12:12
Why is that better than passing "$DIST/updates"?
12:12
<vagrantc>
alkisg: ah, it also doesn't handle: http://foo.bar/debian BAZ/
12:12
<johnny>
debian is always causing troule
12:12
/me blames vagrantc
12:12
<vagrantc>
alkisg: a mirror without components.
12:13rhodan has joined #ltsp
12:13
<alkisg>
vagrantc: and is that case handled by the old code? I can't see it..
12:13
<vagrantc>
johnny: heh. debian invented apt repositories well before i came on board.
12:13
<johnny>
vagrantc, sure.. but you were born..
12:13
so it must have been your fault :)
12:13
<vagrantc>
alkisg: yes, it's definitely handled by the code, as i used it routinely.
12:14
alkisg: though i don't qute remember how...
12:14
<alkisg>
OK, I see it, it checks that with awk
12:15
<vagrantc>
alkisg: it doesn't always add components.
12:15
<alkisg>
vagrantc: why should we be able to call that function with missing parameters?
12:15
What's to gain?
12:15
<vagrantc>
alkisg: they're not missing, they're valid apt sources.
12:16
<alkisg>
vagrantc: I mean, instead of calling with $2=security, call it with the full line: $1="http://security.us.debian.org/ lenny/updates main contrib non-free"
12:16
That "string/flag" thing is not quite clear, I'm afraid..
12:17
<vagrantc>
alkisg: sure. it's some ugly code i got working long ago and it never really broke...
12:17scottmaccal has quit IRC
12:19
<alkisg>
So the last line would be: add_mirror "$SECURITY_MIRROR" with no parameters, but SECURITY_MIRROR would be always properly fixed by 010-security-mirror to be the whole line, "http://security.us.debian.org/ lenny/updates main contrib non-free"
12:19
<vagrantc>
alkisg: sure.
12:21
alkisg: well, i think we should just move the a shared function outside of the mirror handling code, and still keep our divergence in the manage-mirror plugin.
12:22* alkisg was only interested in being able to separate mirrors with enters, for more readable ltsp-build-client.conf - so gladly leaves vagrantc in full charge of any other changes :D
12:22
<vagrantc>
alkisg: you've already got that, and are still proposing more :P
12:23tstafford has quit IRC
12:23
<alkisg>
vagrantc: nope, I didn't - it was YOUR idea to merge the 2 functions, I just offered to help!
12:23
<vagrantc>
i don't believe it was.
12:23
<alkisg>
(and I'm still offering, as long as I understand what should be done)
12:23
<vagrantc>
at any rate, here we are...
12:24
<alkisg>
(09:47:19 πμ) vagrantc: alkisg: that looks much simpler :)
12:24
(09:47:25 πμ) alkisg: Thanks :) <== satisfied
12:24
(09:50:10 πμ) vagrantc: we could probably sync more of the code between those two functions rather than fork it... <== wants more :D
12:24
:P
12:25
<vagrantc>
alkisg: i think i was referring to functions
12:25
er, plugins
12:25tstafford has joined #ltsp
12:26
<alkisg>
Heh, anyway, I don't think it matters :) If you want any help, I'll gladly do whatever I can.
12:26
<vagrantc>
well, i always try to minimize the diff between ubuntu and debian, and here we have a bunch of unnecessary diff.
12:26hersonls has quit IRC
12:27
<vagrantc>
seems like ubuntu doesn't support specifying the whole line for security and updates mirrors...
12:27hersonls has joined #ltsp
12:28
<vagrantc>
but it wouldn't be hard to fix that.
12:29
oh, it looks like it's already handled int he security mirror plugin.
12:31
hrm. seems like the code in the function is completely unecessary.
12:32
<alkisg>
vagrantc: so the differences in ubuntu/debian are how they would "normalize" security mirrors? Why wouldn't *that* be in a different function then?
12:32
<vagrantc>
alkisg: or plugin...
12:32tstafford has quit IRC
12:32
<vagrantc>
alkisg: the other difference is the updates mirror, which debian doesn't have.
12:33
alkisg: so that would have to be a separate plugin somehow.
12:33
if we really wanted to sync them
12:34* alkisg is afraid he doesn't yet know enough about mirrors to be of much help here... :-/
12:36
<vagrantc>
alkisg: seems like you know plenty to me... :)
12:36
<alkisg>
I can code, but I don't understand the problem well enough to code it
12:36
E.g. I don't even know why there's only ONE security mirror, instead of allowing multiple...
12:38
<vagrantc>
ah.
12:38
well, i'd say security-mirror is just a special-cased extra-mirror that has a little different defaults.
12:38litlebuda has joined #ltsp
12:38
<vagrantc>
i.e. mangling the $DIST appropriately. same for updates-mirror
12:39etyack has quit IRC
12:39rhodan has quit IRC
12:40rhodan has joined #ltsp
12:41* alkisg is getting closer, but not there yet... scratches head... :D
12:43
<alkisg>
(thinking also about translators) maybe we could have different plugins for each different type of mirrors?
12:43
This way, we'd have 2 different plugins, for security and updates... the rest would be common
12:43
<vagrantc>
alkisg: yeah, that's the direction my brain is heading...
12:43
<alkisg>
...and the translators would only need to translate the strings once
12:43
<vagrantc>
alkisg: the translators already only need to translate the string once.
12:44
it's smart enough to combine identical strings.
12:44
<alkisg>
Really? Hmmm then the duplicates were because of casing problems, I see...
12:44
<vagrantc>
right
12:45
it's looking much better after i that rash of commits i made.
12:45
i think it went from 95 strings down to 90 just by matching up all the distro's case and removed trailing "."
12:46spectra has joined #ltsp
12:47
<vagrantc>
i guess only 6 of the 90 strings are ALTLinux specific...
12:49* alkisg notes that security comes after the extra mirror on debian, but it's the opposite on ubuntu... does that matter?
12:50
<vagrantc>
alkisg: only slightly.
12:51
<alkisg>
OK, then we could have 010-manage-mirrors in common, and put all distro specific code in the security mirror and updates mirror plugins...
12:51
<vagrantc>
alkisg: the order of entries in sources.list determines where a package will come from ... there's only like 3 security mirrors for debian, so if the newest version is available in another mirror, it's better to pull from one of those other mirrors
12:51
<alkisg>
If those come "later", they'll even "find" the add_mirrors function and be able to use it
12:51
<vagrantc>
alkisg: yup. that's what i was thinking.
12:52
alkisg: i think we should just put the add_mirrors function into /usr/share/ltsp/ltsp-vendor-functions
12:52
alkisg: then it's just there.
12:54
alkisg: my big wonder is do we set up compatibility functions? i.e. if someone made a custom plugin that uses those functions...
12:54
<alkisg>
Oooh
12:54
<vagrantc>
oh do we just inform that they're gone...
12:55
<alkisg>
Ehm, I don't think we're exposing any APIs, so... :D Still, we could rename the function so that he notices
12:55* vagrantc checks the debian-edu codebase...
12:56
<vagrantc>
looks good to me.
12:59
<alkisg>
Btw, here's a 030-blacklisted-packages I made for personal use (to prevent networkmanager, edubuntu-artwork etc from being installed), do you think it also belongs upstream? http://ltsp.pastebin.com/m365538cf
13:02etyack has joined #ltsp
13:03
<vagrantc>
alkisg: ah, curious.
13:03* vagrantc wishes apt pinning was more flexible.
13:04* alkisg picked the example from somewhere on the net and just verified that it works :D
13:10
<vagrantc>
it's amazing how much cleanup one has to do before just starting to work on some code changes.
13:22
hmmm... i'm debating weather to add functions that are vendor-specific to ltsp-build-client, or weather to just use the ltsp-vendor-functions available in ltsp-common-functions...
13:23
the latter effectively adds an ltsp-build-client specific function to all server-side scripts that source ltsp-common-functions... which isn't particularly clean.
13:36
alkisg: ok, layed the ground work for ubuntu and debian to share more...
13:37
<alkisg>
<newbie-mode>vagrantc: so one of the *cough* Ubuntu guys may now see if he can do the same, so that we have a common 010-manage-mirrors?</newbie-mode>
13:38* alkisg has to go, will have a look at that tomorrow. Bye all...
13:39
<vagrantc>
alkisg: or you or i could just commit it :)
13:39
alkisg: now i can more easily insert your change...
13:39alkisg has quit IRC
13:53pmatulis has quit IRC
13:59
<_UsUrPeR_>
hey all. Does anybody know a lts.conf command to have a client look for a specific port?
14:01
<johnny>
a specific port for what?
14:02
<_UsUrPeR_>
whoops. wrong question. I got it.
14:03mushroomtwo has joined #ltsp
14:04mushroomtwo has joined #ltsp
14:05bobby_C has joined #ltsp
14:06hersonls has quit IRC
14:10japerry has joined #ltsp
14:17mushroomblue has quit IRC
14:18hersonls has joined #ltsp
14:41
<vagrantc>
stgraber: well, i'm hoping my last commit isn't resulting in disaster for you... but i couldn't resist incorporating alkisg's fancy new mirror handling code.
14:41
:)
14:51Lns has joined #ltsp
14:53cliebow has quit IRC
14:59spectra has quit IRC
15:00etyack has quit IRC
15:04etyack has joined #ltsp
15:17dro has quit IRC
15:17shawnp0wers has joined #ltsp
15:18
<stgraber>
vagrantc: we'll see ;)
15:30etyack has quit IRC
15:32jcastro has quit IRC
15:33jcastro has joined #ltsp
15:34jcastro has quit IRC
15:34Lns has quit IRC
15:35jcastro has joined #ltsp
15:35Lns has joined #ltsp
15:37jcastro has quit IRC
15:38jcastro has joined #ltsp
15:39F-GT has quit IRC
15:40jcastro has joined #ltsp
15:43hersonls has quit IRC
15:44hersonls has joined #ltsp
15:49
<vagrantc>
ogra: did compcache ever get anywhere in ubuntu?
15:50
seems like it's just now making it's way to debian...
15:50
stgraber: or if you know...
15:51
<prpplague>
/nick prpplague_afk
15:51
oops
15:51prpplague is now known as prpplague_afk
15:54
<stgraber>
vagrantc: it's on by default and provides some swap to all ubuntu thin clients
15:55hersonls has quit IRC
15:55
<vagrantc>
stgraber: oh, nice :)
15:56japerry has quit IRC
15:58shawnp0wers has quit IRC
16:15CAN-o-SPAM has quit IRC
16:35
<vagrantc>
"while read" is amazing.
16:40etyack has joined #ltsp
16:40wwx has quit IRC
16:42wwx has joined #ltsp
16:51johnny has left #ltsp
16:57* vagrantc hopes all this code works better
17:09
<vagrantc>
well, it sure works faster... but i broke something.
17:13
that'll teach me to commit without testing...
17:13johnny has joined #ltsp
17:14
<vagrantc>
i wonder if setting IFS is messing with it somehow...
17:15
ah, yes. it almost certainly is.
20:33ltspbot has joined #ltsp
20:33
<sbalneav>
WAKE UP BOT
20:33
!docs
20:33
<ltspbot>
sbalneav: "docs" :: For the most current documentation, see https://sourceforge.net/apps/mediawiki/ltsp/index.php?title=Ltsp_LtspDocumentationUpstream
20:33
<sbalneav>
^^^
20:35
<comfrey>
thanks sbalneav, looks good
20:36
looks like it is missing some "tricks" that the wiki had.
20:36
but seems like a good document
20:37
i wonder if anyone can comment as to the stability of atom boxes as thin clients...
20:37
vagrantc mentioned some stability issues in the past
20:38
<sbalneav>
All the tricks on the wiki were for 4.2
20:38
<comfrey>
good point sbalneav
20:38
<sbalneav>
That's the difinitive document for ltsp5
20:38
<comfrey>
ok, that makes me feel good
20:39
<vagrantc>
comfrey: haven't heard anything
20:39
comfrey: re: atom boxen
20:39
the docs are even available in debian itself, now. :)
20:40
comfrey: so straight debian lenny, no backports or nothin, eh? :)
20:40
even i barely remember how to use that old thing.
20:41
<comfrey>
there is a good backport then?
20:41
<sbalneav>
I've got ten atom workstations at work.
20:42
<comfrey>
i don't see and ltsp lenny-backports...
20:42
but i will do whatever you recommend vagrantc
20:43
i want to get the classroom thingy going
20:46
<vagrantc>
comfrey: there's nothing at backports.org just yet, though should be soon
20:46* vagrantc nudges Ryan52
20:47
<comfrey>
soon sounds positive :)
20:47
<vagrantc>
comfrey: though i wouldn't recommend using the backports unless you actually need a feature from them...
20:47
<comfrey>
ok , well seems like some research is in order
20:47
<vagrantc>
comfrey: until recently, all the packages in squeeze could be installed on lenny without changes.
20:48
so i didn't really do much work with backporting
20:48
but now they actually need to be rebuild
20:48
but freegeek's running the latest on lenny, so i know at least most of the features work.
20:49
<comfrey>
cool, i'll check the changelogs and see what trinkets await
20:49
<vagrantc>
the main feature i'd say is localapps/fatclients
20:50
<comfrey>
yeah, i have seen some fat client talk here lately,
20:51
<vagrantc>
there's probably something else i'm forgetting, though
20:52
<Ryan52>
vagrantc: hi
20:52
vagrantc: what?
20:52
vagrantc: if it's about backports, I never saw you give me a bzr branch or a .dsc...
20:53
<vagrantc>
Ryan52: ah, i just mentioned it in irc... should i email you?
20:53
<Ryan52>
I should respond to IRC.
20:54
hmmm..
20:54
<vagrantc>
Ryan52: http://llama.freegeek.org/~vagrant/debian/lenny-backports/
20:54
<Ryan52>
and I did completely miss it.
20:54
<vagrantc>
Ryan52: ldm and ltsp are ready now. ltspfs hasn't yet hit testing, so it'll have to wait till then.
20:55
well, ltspfs at that version.
20:55litlebuda has quit IRC
20:55
<vagrantc>
Ryan52: i guess that's 5 days out...
20:56
<Ryan52>
vagrantc: add fixed versions to this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=565962
20:56
<vagrantc>
Ryan52: ah, true. it fixes a security bug...
20:57
Ryan52: so they all should be acceptible for an upload.
20:57
<Ryan52>
no, you still can't do it with that version number.
20:57
<vagrantc>
Ryan52: ?
20:58
<Ryan52>
for the same reason as why you can't upload anything that's not in testing...
20:58
to bpo
20:58
are we talking about the same thing? I interpreted "so they all should be acceptible for an upload." as meaning an upload to backports.
20:58
<vagrantc>
Ryan52: yes.
20:58
Ryan52: the version in unstable has the bug you referenced fixed.
20:58
<Ryan52>
I know...
20:58
<vagrantc>
Ryan52: the versions in testing and unstable still have the security bug ...
20:59
<Ryan52>
vagrantc: why didn't you upload with urgency=medium?
20:59
<vagrantc>
er, testing/stable
20:59
<Ryan52>
yes, so add fixed versions but don't close the bug...
20:59
anyways.
20:59
<vagrantc>
Ryan52: because i filed the bug a few days afterwards, realizing it was a security issue.
20:59
<Ryan52>
ah
21:00
<vagrantc>
but yeah, i should mention the fixed versions.
21:00
<Ryan52>
vagrantc: so set the fixed versions, ask the release team to change it's urgency to medium, it'll migrate at next britney run, then we can upload to bpo.
21:00
<vagrantc>
Ryan52: fair enough.
21:01
Ryan52: ldm and ltsp don't actually require the newer version of ltspfs, so they could be uploaded immediately...
21:02
<Ryan52>
ya
21:03
I'll look into it soon, tho for now I was supposed to be somewhere 3 minutes ago, and I'm getting there by bike, so...gotta go. :P
21:04
<vagrantc>
Ryan52: fun ride
21:07* vagrantc returns to breaking upstream
21:37johnny has left #ltsp
21:56
<vagrantc>
ok, this time i think i won't break it.
22:17johnny has joined #ltsp
22:18try2free has joined #ltsp
22:21try2free has left #ltsp
22:29etyack has quit IRC
22:31etyack has joined #ltsp
22:47cyberorg has quit IRC
22:57
<vagrantc>
that was an exhausting round of breaking shit
22:57cyberorg has joined #ltsp
22:58alkisg has joined #ltsp
22:59
<vagrantc>
alkisg: hi!
22:59
<alkisg>
Good morning! Hey, you did a lot of work yesterday, thanks!
22:59etyack has quit IRC
23:00
<vagrantc>
alkisg: i spent about half of my time trying to figure out why it was broken, only to realize that i hadn't applied the necessary patch for debian to test for writeable /etc
23:01
<alkisg>
:-/
23:01
<vagrantc>
but the good news is that i finally figured it out
23:02
the code for login with localapps enabled is much faster, and *maybe* easier to read... though i'm too delerious to judge that at the moment.
23:03
alkisg: apologies for not crediting you properly... i grabbed the code you emailed me and tweaked it slightly...
23:03
<alkisg>
Don't worry about such things at all, I only care that it works :)
23:04
<vagrantc>
heh
23:04
<alkisg>
Has anyone actually timed how long does that user/group parsing take?
23:05
<vagrantc>
i did some inacurate perception based timing with the old code and the current code...
23:05
on my virtual machine on my laptop, approximately 14-18 seconds with the old code from hitting the login button ... the new code is 8-10 seconds.
23:05
<alkisg>
Oooooh
23:06
<vagrantc>
i think NFS is considerably slower... so it might not be as obvious with NBD.
23:06
<alkisg>
I'm having some big delays when logging in (noticed with an atom fat client) and I was wondering if ltsp was to blame, or gnome, or ...
23:06
<vagrantc>
the main improvement was switching to a case statement that parsed all of the groups in a single pass, instead of a for loop which processed them one at a time.
23:07
well, not a single pass, but not as nested for loops which get longer with each iteration
23:07
<alkisg>
I think that will help a lot in nbd, too...
23:07
<vagrantc>
it certainly won't slow it down...
23:07
<alkisg>
\o/ for vagrantc :)
23:08
<vagrantc>
i wish i had figured out my false negatives sooner...
23:08
nested case statements are a little ugly looking
23:09
<alkisg>
This is really a work for some other language, though... maybe python
23:09
<vagrantc>
yes.
23:09
python has a pretty harsh startup cost, though... so it'd have to be handled in a single python script called only once for it to really matter.
23:09
<johnny>
i wouldn't be upset if ltsp-buiild-client was converted to python..
23:10
and the plugins
23:10
<alkisg>
johnny: I'm not sure about that, as they're too much command line oriented
23:10
<vagrantc>
i tried implementing part of the code i messed with today in python, but did so with multiple calls to python... and it wasn't worth it.
23:10
<johnny>
it could be more api oriented.. and some wrappers
23:10
vagrantc, perhaps running python as a server or somethin
23:10
so it only starts up once
23:10otavio has quit IRC
23:11
<johnny>
i wonder if anybody has ever done that
23:11
<vagrantc>
johnny: that's a great way to consume an absurd amount of memory.
23:11
at least from a thin-client perspective
23:12* alkisg needs to get ready to go skiing... goodbye all
23:13alkisg has quit IRC
23:13
<johnny>
vagrantc, just the mem for a single python process?
23:13
seems like it shouldn't be too crazy..
23:13
<vagrantc>
johnny: well, this is python we're talking about
23:13
johnny: how many MB sound reasonable to you?
23:13
<johnny>
4? 5?
23:13
depending on how much you're doing with it..
23:14
for just one operation it's too much
23:14
but if everything is using it..
23:14
of course.. if the chroot used a different c lib.. it'd probably be quite a bit smaller
23:14
<vagrantc>
that's a pretty steep spike on even a 128MB client
23:15
<johnny>
vagrantc, anybody done experiments at using a different C lib?
23:15
vagrantc, i wonder about using vala
23:16
vala might be better than python
23:16
and we already dep on glib for ldm
23:16
<vagrantc>
may as well write something in C
23:19
<johnny>
except vala is easier to write
23:19
and doesn't need to be compiled
23:19
iirc anyways
23:20
that reminds me..
23:20
i want a vim.. but with a better configuration syntax..
23:20
perhaps based on glib..
23:20
and extendable with vala or gjs
23:21
<vagrantc>
well, i'm spent too many consecutive hours on ltsp today... made good progress... but i'm tired.
23:21vagrantc has quit IRC
23:45cyberorg has quit IRC
23:45cyberorg has joined #ltsp