**** BEGIN LOGGING AT Mon Jun 26 20:43:56 2006

jun 26 20:48:16 *       Bman^ (~andre@193.90.60.81) has joined #debian-edu
jun 26 20:54:22 *       Channel #debian-edu modes: +nt 
jun 26 20:54:22 *       Channel #debian-edu created on Mon May 08 11:22:00 2006
jun 26 20:58:33 *       knuty has joined
jun 26 20:58:36 *       h01ger waves 
jun 26 20:59:05 *       k4x = Kurt Gramlich
jun 26 21:00:47 *       h01ger tears his hair with tuxtype. after the licence mess i now can convert it to utf8 before i can/want to release it...
jun 26 21:03:29 sepski  who's running the show ? 
jun 26 21:03:42 *       sepski = Ronny Aasen
jun 26 21:04:24 *       xorAxAx = Alexander Schremmer
jun 26 21:04:38 xorAxAx now the southern americans know skolelinux
jun 26 21:04:54 xorAxAx i wonder if nobody has used it with locale es? it has a show blocking bug :)
jun 26 21:05:00 k4x     xorAxAx: congrats ;-)
jun 26 21:05:06 xorAxAx but i didnt tell anybody about it :)
jun 26 21:07:50 sepski  are there no moderator for this meeting ? should i just talk about partman and be done with it ? :)
jun 26 21:09:00 h01ger  shouldnt we follow the agenda?
jun 26 21:09:01 h01ger  :)
jun 26 21:09:50 sepski  aye, but usualy someone calls to order or some moderatoish buisiniss.... im just a tech i have no social skills :) 
jun 26 21:10:52 h01ger   /topic 1. Status of Autopartitioning   1.1 partman 1.2 autopartkit
jun 26 21:11:08 h01ger  i cant set the topic, cause i'm not channel operator :(
jun 26 21:12:05 sepski  well i'll just say what i have to say about partman then. 
jun 26 21:12:18 knuty   sepski: do that :)
jun 26 21:12:40 sepski  first of all. partman work it's preseedable (with a few exceptions) 
jun 26 21:13:49 sepski  without lvm it is down to 1 question, the confirmation dialog. with lvm we need 2, but that's also fixed in the next partman-auto-lvm http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=375265
jun 26 21:14:34 sepski  currently we use partman/9, the fix requered for preseeding with lvm is in partman/13, the good news is that it's soon going to be a beta3 installer and hopefully its coming into etch then.
jun 26 21:15:13 knuty   sepski: Thats really good news.
jun 26 21:15:15 sepski  when that happen partitioning is done in 1 question, the confirmation dialog
jun 26 21:15:41 sepski  another issue that's not so good is that partman can only make 1 (one) volume group
jun 26 21:16:11 sepski  the name of the volumegroup is made as a preseed and not tru the recipe
jun 26 21:16:21 *       h01ger wondered today why we build against the etch installer, while the d-i team builds against sid 
jun 26 21:16:58 sepski  h01ger, good question i think becouse it takes work to convert to sid's
jun 26 21:17:12 sepski  that leave us with a few options. 
jun 26 21:17:58 h01ger  sepski, how much do we differ from normal d-i? (briefly, we can discuss details after the meeting..)
jun 26 21:18:20 sepski  we can 1: use 1 vg and not worry about having data and system lv's in the same vg, 2: we can reserve space for vg_data in the recipes and make vg_date and lv's in the cfengine run or similar, or rewrite partman to take vg config tru the recipes
jun 26 21:18:33 xorAxAx h01ger: they build against sid for the etch images?
jun 26 21:19:04 sepski  h01ger, xorAxAx they develop on sid and have daily images
jun 26 21:19:16 sepski  they are soon making a beta3 installer. 
jun 26 21:19:33 h01ger  xorAxAx, no. the daily-builds are against sid. etch builds are seldom. (beta3 is one)
jun 26 21:19:36 sepski  when the installer freeze we should build against that. but now it would be better to build against sid's :)
jun 26 21:19:56 xorAxAx sepski: yeah, right ... so they will migrate their packages before the next beta?
jun 26 21:20:06 sepski  any thoughts about what to do with volumegroups ? 
jun 26 21:20:27 sepski  rewriting partman is way out of my leauge. and i dont think anyone else have that much time either
jun 26 21:20:32 h01ger  xorAxAx, yes
jun 26 21:22:05 sepski  putting everything in a single vg is the easy way out, then it works as it  is now. making room for a partition in the recipes and making the vg later is a posibility, but will require some work
jun 26 21:22:43 sepski  if there is no opinions on it i will continue
jun 26 21:22:50 knuty   What is the recomend way to solve volume groupe issue other than placing everything on one vg
jun 26 21:22:52 knuty   ?
jun 26 21:23:02 h01ger  do it manually
jun 26 21:23:10 h01ger  its not preseedable _AIUI_
jun 26 21:23:41 knuty   h01ger: So we has to convice the d-i partman people?
jun 26 21:23:54 knuty   To make support for 1+n volume groups?
jun 26 21:24:06 h01ger  they know its a nice feature :) 
jun 26 21:24:15 sepski  they have too much on their time already before the installer freeze, they will use out patch if we make it. 
jun 26 21:25:05 h01ger  sepski, #375265 is fixed. so the problem left (1+n volume groups) is the same as preseeding multiple discs, right?
jun 26 21:25:18 sepski  would it be realy bad to have everything in a single volumegroup ?
jun 26 21:26:25 sepski  that bug was just to have the ability to preseed the what disk to partition question ,and get it with lvm
jun 26 21:26:43 sepski  partman-auto can only handle 1 disc atm afaik
jun 26 21:26:56 knuty   sepski: From a maintainability standpoint i believe not, but that could be changed later. Now the installer should work to get other things done. 
jun 26 21:27:04 *       h01ger dont think its really bad, its what we have in sarge :) more would be better, obvioulsy
jun 26 21:27:15 sepski  then ill leave it as 1 vg and we look at what can be done later. 
jun 26 21:27:26 sepski  now another issue. 
jun 26 21:27:48 knuty   sepski: Does it crash. If partman-auto can handle on disk don't touching two disks or make a RAID it's ok for developer purposes.
jun 26 21:28:00 knuty   sepski: Yes the other issus
jun 26 21:28:53 sepski  if there is a vg with the same name as the one beeing created on the disc already the lvm tools recognice the signature and refuse to overwrite it. this is checked for in the installer to prevent a bad crash, but it will prevent you from partitioning the disc. 
jun 26 21:29:21 sepski  the bug http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=347479
jun 26 21:30:14 sepski  it's not realy much we can do anything about. i guess we could use a different vg name, but i dont think that's to clever either.
jun 26 21:31:52 h01ger  so are we done with the partman topic? :)
jun 26 21:31:56 sepski  nope :)
jun 26 21:31:59 h01ger  or are there more issues?
jun 26 21:32:15 sepski  the recipes need more work. 
jun 26 21:32:20 sepski  and another. 
jun 26 21:32:38 sepski  autopartkit seam to have configs for tmpfs filesystesms ?? whats that about ? 
jun 26 21:32:43 sepski  just a clever way to feed fstab ? 
jun 26 21:33:05 h01ger  sepski, where/what configs? 
jun 26 21:33:54 sepski  line 8 of debian-edu-install/autopartkit/Main-Server.table
jun 26 21:34:22 sepski  i was looking at autopartkit tables to get min and max values for partition sizes when i noticed it
jun 26 21:35:56 *       h01ger thinks its a clever way to feed fstab
jun 26 21:35:59 sepski  but i guess that can be done as part of the cfengine run rather easily. 
jun 26 21:36:26 sepski  then i was about done with the partman part of the installer. 
jun 26 21:36:51 sepski  generaly it works a few minor issues and i think it will work rater well. 
jun 26 21:38:12 sepski  keep in mind that the installer today makes non lvm partitions ill change that when we have partman-auto-lvm/13, but is very easy to change
jun 26 21:38:37 h01ger  sepski, are u in touch with Alphix? - he seems to work on lvm autopartioning in d-i...
jun 26 21:39:23 sepski  h01ger, i have asked a few questions in debian-boot, but it's usualy  Frans Pop that answers me
jun 26 21:39:44 h01ger  :)
jun 26 21:39:45 sepski  he also fixed the preseed + lvm issue that's coming in partman/13 now
jun 26 21:39:56 sepski  (in just 1 day) :))
jun 26 21:40:49 h01ger  the mailinglist is also nice to have an overview, bugs and bugs-done are posted there as well :)
jun 26 21:41:09 sepski  autopart goes out of the agenda since white is afk. he wrote a email about it. 
jun 26 21:41:26 h01ger   /topic "2. the status of the installer" is also done or is there something else?
jun 26 21:41:33 sepski  no 
jun 26 21:41:40 sepski  i mean yes
jun 26 21:41:42 sepski  there are more
jun 26 21:42:13 knuty   sepski: ok
jun 26 21:42:16 sepski  i have only tried 1 compleate install lately, but that failed miserably. seams like atlest part of the cfengine run failed. 
jun 26 21:42:42 sepski  squid is in a nonstartable fashion atlest there is no network config and no hostname. cousing most other services to fail as well
jun 26 21:43:14 sepski  i have not investigated it, but i had hoped other's had tested it more then i have.
jun 26 21:44:03 sepski  i do know that there was no ip on the interface when the installer was running. thay have a bad effect on the cfengine run atleast. 
jun 26 21:44:25 sepski  have anyone else tried etch-test ? 
jun 26 21:45:43 sepski  .oO(i had hoped it only was a local issue)
jun 26 21:47:14 *       h01ger thinks noone who is awake has tried it recently
jun 26 21:47:27 sepski  well they can read about it in the irc log then
jun 26 21:47:35 *       h01ger nods
jun 26 21:47:50 sepski  ill write a bug when i have time to test it again
jun 26 21:48:02 knuty   I think we should use 15 minutes on the rest of the issues
jun 26 21:48:04 sepski  i have nothing else on the installer
jun 26 21:48:28 knuty   The idea to fix the partitioning is that we can install the CD and fix the other bugs
jun 26 21:49:23 h01ger  this also can be done by doing sarge installs and upgrades :)
jun 26 21:49:25 sepski  well partitioning work now. with 1 questiong giving you a working non lvm'd system so its free to test for all
jun 26 21:49:53 knuty   sepski and the other developers: Well done!
jun 26 21:49:55 sepski  depending on the recipe you pick since not all are finished
jun 26 21:50:09 knuty   sepski: What about bug 311188
jun 26 21:50:10 sepski  but main is okayish and ltsp is close
jun 26 21:50:31 h01ger  sepski, maybe write a mail for people wanting to test about how to workaround the problem
jun 26 21:50:44 h01ger   /topic 3. #311188
jun 26 21:50:49 sepski  h01ger, ill write a call for testers email
jun 26 21:51:01 h01ger  sepski, great! :)
jun 26 21:51:53 knuty   Could we go to /topic 4 if /topic 3 was abou sepski sending e-mail to call for testes?
jun 26 21:52:02 h01ger  about #311188: no news afaik, but read the bugreport if you haven't done yet :-) my plan after finishing NM (which includes a tuxtype upload..) is to work on the bugs blocking #311188, and i would be happy about help :)
jun 26 21:52:02 sepski  i have not been much involved in bug 311188, i had hoped it would be as easy as each package renaming it's conffile from eg: squid.conf to squid.conf.orig and symlinking to squid.conf in postinst. then we coulr change that symlin to squid.debian-edu.conf and be policy compliant
jun 26 21:52:29 sepski  but i dont know if that would actualy make us policy compiant since IANAL
jun 26 21:52:52 h01ger  IANADD too :-D but i dont think that would work :)
jun 26 21:53:48 sepski  pity. it would have been very easy dor both us and the package devs
jun 26 21:53:57 h01ger  some maintainers have replied to the bugs and provided insightful answers, so my idea is to fix bug by bug (and adopt debian-edu-config also), until 311188 is fixed
jun 26 21:54:25 h01ger  it would also make upgrades a greater pain ;-)
jun 26 21:54:25 sepski  h01ger, good, ill help where i can. 
jun 26 21:54:35 h01ger  great!
jun 26 21:54:36 sepski  h01ger, ah true. 
jun 26 21:54:46 h01ger  so next point?
jun 26 21:55:03 sepski  4.One Laptop per Child developer board
jun 26 21:55:23 h01ger  knuty, ?
jun 26 21:55:39 knuty   Ok. In short, it seems that we get Extremadura on board and ...
jun 26 21:56:03 knuty   and Christian Perrier. 
jun 26 21:56:31 knuty   I did not see any rason to apply for a OLPC developer board if nobody want
jun 26 21:56:32 knuty   it
jun 26 21:56:44 *       h01ger is interested in OLPC but too busy atm
jun 26 21:57:13 knuty   So no people want to join that brance, and I can order boards, also to you h01ger :)
jun 26 21:57:55 sepski  sorry for beeing ignorant ??? this is to high level english for me... iis this board like a council ?   company board ? 
jun 26 21:58:08 knuty   So we will apply for enough board to people that does real work. I will also establish a Trolltech qt Community student project 
jun 26 21:58:10 h01ger  no, a prototype board
jun 26 21:58:13 sepski  oohhh
jun 26 21:58:16 h01ger  not a laptop yet :)
jun 26 21:58:20 h01ger  without display
jun 26 21:58:21 sepski  <-- *smach* stupid me
jun 26 21:58:29 knuty   sepski: It's ok
jun 26 21:58:50 sepski  if it's about getting the debian-edu installer to work on it i probably find some time
jun 26 21:59:48 knuty   To have a joint development with Extremadura will really be a good thing. Getting student projects on this is also good, and I can continue with some Skolelinux work in Trolltech as a qt Community Manager
jun 26 21:59:55 h01ger  sepski, exactly thats the challange :)
jun 26 22:00:16 xorAxAx knuty: you mean the OLPC would be the reason to join dev?
jun 26 22:00:35 sepski  if i can get one i can try sneaking it past the fiancee's no more computer junk curfeew
jun 26 22:00:39 sepski  :)
jun 26 22:00:41 knuty   xorAxAx: dev email list?
jun 26 22:01:10 xorAxAx knuty: sorry, i mean to have " joint development"
jun 26 22:01:38 knuty   xorAxAx: Jepp. It's a way to make different people work togheter on a common goal to our common benefit
jun 26 22:02:12 sepski  we are over time we should speed up
jun 26 22:02:18 xorAxAx thats right. i am not sure if there might be other goals first that make team work easier
jun 26 22:03:03 knuty   I'll change to /topic 5 if it's ok
jun 26 22:03:04 *       h01ger thinks OLPC has a political vision and so can get a different/another dynamic. 
jun 26 22:03:22 *       h01ger also thinks "next point"? 
jun 26 22:03:46 sepski  ack
jun 26 22:03:58 knuty   me /topic 5 applying for travel expenses for making the Minverva applications
jun 26 22:04:25 *       Werner == Morten Werner Olsen (tooooo late, sorry).
jun 26 22:04:58 k4x     for germans the paper is prepared, i need still more information fro, Thierry
jun 26 22:05:23 knuty   Oh you should call your national Minerva/Socrates office http://wiki.debian.org/DebianEdu/DevCamp2006/ApplicationLetter
jun 26 22:05:30 k4x     at the moment i am not sure, how many germans will join the meeting, we are still collecting
jun 26 22:05:49 xorAxAx k4x: its important to submit it on time - when is the exact deadline?
jun 26 22:05:54 k4x     knuty: yes, i know, the info is in the debian wiki
jun 26 22:05:55 knuty   k4x: thats ok, but you should send the letter this week
jun 26 22:05:56 Werner  it's "only" travel expences that the Minerva project covers, right?
jun 26 22:06:21 xorAxAx Werner: no, but this is pre-minerva-application from what i gathered
jun 26 22:06:31 k4x     knuty: i will send it this week
jun 26 22:06:39 knuty   Werner: The Norwegian person at the Minerva Office told me that they chould cover expenses up to 1000 EURO for each person from our contry
jun 26 22:07:20 Werner  knuty: that will also cover lost work for people without vacation? :)
jun 26 22:07:25 knuty   Thats a big if, If we write a acceptable application for money, and taking into concideration that EU is changing the Minerva program
jun 26 22:07:49 Werner  ok
jun 26 22:08:00 *       Werner will have to look at the Norwegian application..
jun 26 22:08:35 knuty   Werner: Difficult to say. I did not got ask about that. I'll read the application paper better tonight, and you should check too: 
jun 26 22:08:37 knuty   http://siu.no/vev.nsf/O/SOKRATES-Minerva-Soknad+og+rapportering-Soknadsskjema
jun 26 22:09:14 Werner  thanks
jun 26 22:09:49 xorAxAx note that these travel expenses are outside of the project which should be applied for at the minerva program, right?
jun 26 22:10:47 knuty   xorAxAx: Everybody that travels from somewhere other country than France should apply, and everybody are joning the application work in some way or another.
jun 26 22:11:25 xorAxAx knuty: umm, i mean, this all is pre-minerva-program-application
jun 26 22:11:33 xorAxAx i.e. skolelinux isnt yet in, is it?
jun 26 22:11:43 knuty   So put every german on the application, because we will need input from everyone. I know that we should do that. But dont expect us to get the money
jun 26 22:11:50 *       xorAxAx waits for kurt to answer those 4 questions above
jun 26 22:12:25 xorAxAx knuty: what do you mean by that? esp. by "we will need input"?
jun 26 22:12:49 xorAxAx every legal entity has to apply for travel funds, so this is about different "we"s :)
jun 26 22:12:58 k4x     xorAxAx: exact deadline i do not know, i will phone tomorow
jun 26 22:13:01 knuty   xorAxAx: It's not complicated. Applay for the workshop making a Minerva EU-founded pre-project, and focus on sharing knowledge, open content and cooperation between countries in education and getting people a new chance and so on. 
jun 26 22:13:04 *       xorAxAx needs to know because of train tickets
jun 26 22:13:22 knuty   k4x: smart, call them tomorow.
jun 26 22:13:41 *       Werner sets mode -t #debian-edu
jun 26 22:13:41 *       ChanServ sets mode +t #debian-edu
jun 26 22:13:41 xorAxAx knuty: yes, so funding will be different than in the future when skolelinux is assumed to be in the program
jun 26 22:13:47 xorAxAx hehe
jun 26 22:13:50 xorAxAx Werner: mode lock
jun 26 22:14:43 sepski  Werner, would it be bad to set chanmode -t ? are there much topic spamming in oftc ? sometimes it would be nice for us mere mortals to update the topic too :)
jun 26 22:14:44 k4x     this all is pre-minerva-program-application
jun 26 22:15:05 *       knuty /topic 5 : The germans and the norwegians will have sendt the applications this week?
jun 26 22:15:18 knuty   k4x: Yes? :)
jun 26 22:15:22 k4x     xorAxAx: skolelinux is starting to get financial support with the devcamp meeting
jun 26 22:15:23 *       Werner sets mode -t #debian-edu
jun 26 22:15:23 *       ChanServ sets mode +t #debian-edu
jun 26 22:15:37 Wedge Werner 
jun 26 22:16:24 xorAxAx k4x: yeah, the upcoming question was if the in-program sponsoring is different than the one for dev camp
jun 26 22:16:28 k4x     knuty: yes, i will send it this week, every german who wants to join should put his name into the wiki page
jun 26 22:16:44 *       knuty /topic 5 closed
jun 26 22:17:08 sepski  topic 6 is  SLX 2.0r1 
jun 26 22:17:15 k4x     xorAxAx: yes it is different afaik
jun 26 22:17:24 *       Werner sets mode -t #debian-edu
jun 26 22:18:47 h01ger  who works on SLX 2.0r1 ?
jun 26 22:19:10 Werner  white: ?
jun 26 22:19:22 sepski  RalfGesellensetter  write that on the agenda, but he also stated he could not be here
jun 26 22:19:48 h01ger  because if its based on sarge_r3, then it also needs a new installer build (new kernel..)
jun 26 22:19:50 Werner  we did some preparations in Extremadura, but I guess white have the details..
jun 26 22:20:20 sepski  i guess we drop this topic until white is online again ? 
jun 26 22:20:24 Werner  yes
jun 26 22:20:44 *       h01ger nods
jun 26 22:20:55 sepski  i guess that concludes the meeting. 
jun 26 22:20:58 h01ger  so the meeting is over :)
jun 26 22:21:10 sepski  20 mins overtime. bah nothing compared to public sector :)
jun 26 22:21:42 Werner  :)
jun 26 22:22:14 *       k4x thanks to all, hope to meet you in France
jun 26 22:22:23 knuty   sepski: the meeting was not started before 19:12 UTC ;)
jun 26 22:23:12 sepski  knuty, true, hard to know who should start it. i assumed frode would since he had so nice an announcement email :)
jun 26 22:23:36 h01ger  sepski, so why dont we build d-i against sid (and make it _install_ etch by default)?
jun 26 22:24:03 sepski  h01ger, i asked white this a few weeks ago. he was reluctant to change the build system
jun 26 22:24:25 Werner  h01ger: but when etch is frozen, we should change back??
jun 26 22:24:27 h01ger  ok, so lets discuss with him
jun 26 22:24:31 sepski  h01ger, then move back to the frozen installer when it freezes
jun 26 22:24:42 sepski  Werner, yes
jun 26 22:25:24 Werner  I guess it's better to help the installer-team to get things into etch, but then we'll lose some valuable test-time on the rest of our installation..
jun 26 22:25:33 h01ger  d-i was branched (for etch) in the (last months of) preparation of sarge, so we would switch to that branch..
jun 26 22:26:18 h01ger  Werner, but doesnt the installed system need more testing than the installation itself? 
jun 26 22:26:18 Werner  h01ger: do you know how we can do this?
jun 26 22:26:35 Werner  h01ger: we need to do a lot of testing, yes..
jun 26 22:26:48 h01ger  Werner, partly. but i can get a clue on the parts i dont know :)
jun 26 22:27:27 sepski  h01ger, part of the problem is debian-cd, we use a old svn version that we patch with our own patch that we dont dare update in case it breaks, and only those with accounts on administrator can do this work. they are also the most busy people :)
jun 26 22:28:27 Werner  didn't white start with a new version of debian-cd??
jun 26 22:29:03 sepski  Werner, yes. but that's a while ago and it's under constant change.
jun 26 22:29:19 h01ger  maybe we need the old version for sarge builds and the new one for etch builds :)
jun 26 22:29:20 Werner  right.. and white doesn't want to update it? :)
jun 26 22:29:33 sepski  Werner, dont think he have had the time to try to 
jun 26 22:29:51 Werner  right.
jun 26 22:30:05 Werner  sepski: and you know the setup a bit ??
jun 26 22:30:37 sepski  Werner, just some. 
jun 26 22:31:10 sepski  Werner, i was trying to help white but i just confused him, since i was working of a different svn version of debian-cd then he did... it got confusing fast :)
jun 26 22:31:44 Werner  right .. but if you want to look more at it, I guess white would welcome some help :)
jun 26 22:31:54 sepski  o got a tarball of his version, but then he already had it working so i didnt look at it any more.
jun 26 22:32:07 Werner  ok
jun 26 22:32:36 sepski  ill ask him if i can get a hold of him in his nondefault TZ down under
jun 26 22:33:58 *       h01ger has changed the topic to: Debian-Edu/Skolelinux 3.0 codename "terra" under development || check services in etch report to  http://bugs.skolelinux.no/1095  || next meeting at 2006-07-xx || stable version is 2.0