=== jussio1 is now known as jussi01 === bddebian2 is now known as bddebian [17:11] morning all [17:14] well, last night I was able to get sabayon to stop crashing at startup when I edit a profile [17:15] LaserJock: how do you apply the profiles made with sabayon? [17:15] alkisg: do you mean in general? [17:16] yeah [17:16] you select a user to belong to a profile [17:16] but .... I haven't gotten that far yet [17:16] In a manual I saw that it was left to the reader.. And I thought, ok for new users, but what if I want to apply some settings to some existing users? [17:16] I'm getting crashers once I get into the profile [17:16] don't read the manual, that's from like 2006 [17:17] Intrepid? [17:17] hmm? [17:17] You get crashes with sabayon on intrepid? [17:17] yeah [17:17] Let me try... [17:18] I fixed one so now I can actually get the session started [17:18] but then I get a ton of database errors and it dies [17:19] to use the profile you click on the profile and then hit the Users button [17:20] then you can pick which users to apply it to [17:20] And how did you bypass the first bug? [17:21] well, it was actually a really little problem where it tries to find the orientation (top, bottom, etc.) of the panels [17:21] for some reason it was calling the function without a panel id and it would then crash [17:21] so I just added a little time that checks if a panel id exists and if it doesn't just assigns it as "top" [17:22] Ah, in the source, it's python, got it [17:22] s/time/code/ [17:22] and it worked ok at first [17:22] but then I got these database permission errors [17:23] looking at the Gnome bugzilla sabayon has 57 open bugs and many of them are crashers [17:23] my feeling is that Sabayon is trying to just do too much, you'd need a real team of people working in it to get it stable/maintained [17:24] this crasher I fixed is probably just because something changed in gnome-panel [17:24] Sabayon interacts so much with the apps that it has to keep up with *their* changes as well [17:25] Yes... it has to be plugin-based, possibly with different maintainers for each plugin [17:26] it's somewhat modular [17:27] but I don't know that right now one can pick which "modules" to use [17:27] Where was this first crash again? In /usr/bin/sabayon or in sabayon-session? [17:28] I can't get past it... [17:28] sabayon-session [17:28] so if you click on "Edit" an Xnest window pops up and it tries to start Gnome [17:29] but default Intrepid for me it just shuts down after a second or 2 and I never see anything in the window [17:29] alkisg: is that what you get? [17:30] Yes, the same [17:32] so now I need to figure out where these DB errors are coming from [17:32] it sort of looks like a permissions problem [17:32] What did you do to fix the first error? [17:33] changed anything in paneldelegate.py? [17:33] http://paste.ubuntu.com/96113/ [17:33] yep [17:34] you want to patch /usr/share/python-support/sabayon/sabayon/sources/paneldelegate.py [17:39] Nah.. I don't even get to see the database crashes [17:39] I see the desktop and then it's killed, but no error logs [17:39] woot, it works for me [17:40] try it again [17:40] I think I had some times where it wouldn't work the first time but if I hit "Edit" again it would [17:42] I tried it some times, most of them plain crashes without error log, on 2 of them I got a dialog to sent a .conf file to the sabayon team (but the file didn't exist) [17:42] ...debugging... [17:44] it's starting up every time here now [17:44] I got something: Failed to connect to socket /tmp/dbus-bkaJq0tpTl: Connection refused [17:44] but it's not saving some changes (removing launchers from the panel) [17:44] hmmm [17:44] Maybe I need to logoff [17:44] hmmm u guys working on sabayon... cool [17:44] I think I saw an upstream bug about that [17:45] nubae: I got it to start up here anyway :-) [17:48] this thing is terribly complex [17:48] yes, is it so complex it would be simpler to re-write? [17:49] that is what i was wondering [17:49] well, not if you want to do what it does [17:49] but it's just tracking *so* much stuff. My simple test profile where I turned off a couple things and edited the menu is a 11MB zip file! [17:50] did u get it to work? [17:50] well, I created it [17:50] nubae: hey, u here today? [17:50] yeah :-) [17:50] I got it working (once at least), but it complained a lot, and didn't save anything [17:52] Hm... it also leaves a lot of processes running after each crash [17:52] May I should kill them before restarting it [17:55] bah [17:55] yeah, kill them [17:55] I think that's what I was running into [17:55] I kept trying and got problems, then this morning I tried it again (without logging out or rebooting) and it works fine [17:56] well, users-admin seems to not be able to add a user! [17:56] Gnome is really falling apart when it comes to basic user management :( [17:56] yep [17:56] I use the command line only now [17:58] not to mention its totally differenta cross all distros [17:58] LaserJock: this happened to hardy as well, users-admin had problems getting root rights didn't complain about it or something [17:59] I see a bug where it looks for user "root" instead of UID 0 [17:59] I wonder if that's it [17:59] I don't think so [18:00] LaserJock: it worked fine for me in hardy [18:01] which is why i think python would be better and modules [18:01] then you could step through it one at a time. i am however NOT a programmer [18:02] well, it doesn't really matter so much what language it's written in [18:02] bottom line there has to be people behind it [18:03] Yes, but if it's interpreter or shell-based, it's easier for admins to debug the problem and send accurate feedback / patches [18:04] guess there is some truth to that [18:04] not every admin is a package maintainer I suppose [18:05] alkisg: I'm not entirely sure about that [18:06] gdb can do a pretty good job of debugging C/C++ apps [18:06] Sure, but does a regular admin that exceriences the problem download the source, configure and compile it, and then run it under gdb to see where the problem is? [18:06] they don't have to [18:07] Why not? How accurate feedback can you give now with the problem in users-admin? [18:07] they might need to install the debugging symbols package, but they don't need to compile or anything like that === jussio1 is now known as jussi01 [18:08] Debuggin on a live system is surely a better way than looking though a crash dump [18:08] yeah [18:09] right, but that doesn't have a ton to do with the language you write it in [18:10] I mean, don't get me wrong, I love Python and prefer it [18:10] but I'm not going to just go rewrite every app that's not in Python :-) [18:10] Of course not!!! [18:11] alright, I found the problem [18:11] users-admin checks that each user has a login, password, home directory, and shell [18:11] I've been an assembly/C/Pascal programmer myself for the last ...ehm... 17 years, so it isn't easy for me to switch to python as well [18:12] Sabayon uses a user that doesn't have a home directory [18:12] so basically if you install Sabayon you can no longer use users-admin :( [18:12] pascal :-) such a educationally popular language [18:12] at least it was [18:12] I never learned Pascal [18:12] and what if u create the /home dir for the sabayon user? [18:13] well, I *think* users-admin would work but I think it would break Sabayon [18:13] Sabyon uses a temporary home dir in /tmp [18:13] * nubae wonders what that would have changed from Hardy [18:14] it worked fine in Hardy with Sabayon installed? [18:14] yes [18:14] I remember adding users and making profiles [18:14] and it working [18:15] I think what happened is the gnome-system-tools-backends changed [18:15] ah, makes sense [18:15] so that it added in these checks [18:15] they probably were fixing a bug where people were creating users without /home dir :-) [18:16] but this is sort of a nasty bug [18:16] there are a lot of packages, not just Sabayon, that create users like this [18:16] ohhhhhhhh [18:17] * LaserJock has a thought [18:17] sabayon creates it's user as a regular user, not a system user [18:17] I bet users-admin on checks for a home dir for non-system users [18:17] *only [18:22] I wonder if anybody uses sabayon in Debian, it's only got 1 open bug [18:23] interesting. i could no longer use users/admin so i simply did a backup and re-install of the os after installing sabyon [18:23] yikes! [18:23] *shrugs*, it worked [18:23] yeah, I reinstall a lot [18:24] I think I'm about the worst of the dev ;-) [18:24] *devs [18:24] I know one guy hasn't installed fresh since like Red Hat 6 or something [18:24] ... and he's running Ubuntu Hardy or newer [18:25] nubae: i'm getting ready to tackle the inet problem if you have time today [18:27] yeah sure, go for it [18:27] paste me /opt/ltsp/fati386/etc/network/interfaces [18:28] so I think if we got a list of "things" we wan to be able to in user management then we can figure out what app(s) are going to be the best to look at [18:29] I think sabayon is too complex to maintain unless you have some people who can spend some serious time on it [18:31] whatcha mean exactly? [18:32] nubae: what did you want pasted? [18:33] nubae: did you mean me? [18:34] LaserJock: yeah [18:34] nubae: well, it seems quite fragile [18:34] Ahmuck ^^ [18:34] it's trying to track *ever* change made to a desktop [18:34] u know i was thinking about the sabyon vs python thingy. being a non programmer it i guess it really doesn't matter. i've heard various things about python that were always positive. i guess from my pov, waiting is the hardest thing, and if someone is volunteering to do it in python and the other application is not being maintained i'm always the person to choose the most eficiant route,... [18:34] ...which is usually the shortest [18:35] and these days there is a lot that's going on [18:35] ah ok, gotcha, so limit what it does? [18:36] well, I'm not sure [18:36] nubae: what language is your script written in ? [18:36] C [18:36] I wonder if it's just not a great overall idea right now [18:36] no :-) bash [18:37] LaserJock: it is [18:37] http://pastebin.be/15745 [18:37] * nubae can't code C to save his life... no patience [18:37] I'm not sure that deployable profiles is really something that Linux can do [18:37] i understand the script, and will prolly customize the apps to my own need [18:38] is there anyway to "link"the system apps with the fat client? [18:38] deployable profiles ? [18:38] that would be like using local-apps [18:38] * Ahmuck-Jr used to code C++ [18:38] 20 years ago [18:38] k, not quite that long [18:38] Sabayon is used to deploy "profiles" [18:39] more like 12 years ago [18:39] Ahmuck-Jr: paste me your ifconfig output [18:39] on the from the server [18:39] I'm just not sure if that's something that Linux/Gnome is going to be able to do very well for the foreseeable future [18:40] well if they stopped thinking about it as being a single user system, then sure [18:40] but I think you're right [18:40] http://pastebin.be/15746 [18:40] well that looks fine, so must be dns not getting through to the thin client [18:41] u got this in /etc/ltsp/dhcpd.conf?: option domain-name-servers 192.168.0.254; [18:41] Even if sabayon worked, I'm not sure I'd want to use it. I'd prefer a tool that would let me specify the changes I want, instead of tracking any change I (even accidentally) made. [18:41] [18:41] do i need to add dns servers to the network [18:42] alkisg: you can manage the changes [18:42] if you want to be able resolve web pages, then yes [18:42] yeah, it just shouldnt track everything, is there an easy way to limit that LaserJock, or is it loads a work? [18:42] LaserJock, inside a 10k-line file? (if I recall correctly the number you've said above) [18:43] nubae: there is some, you can choose what to track and what not to basically [18:44] alkisg: no, that's the source code. The GUI lets you look at changes and tell it to lock/ignore them [18:44] but the problem is that it has to track everything in the first place, to then give you a choice :-) [18:44] http://pastebin.be/15747 [18:44] would limit crashes too [18:44] pessulus is great, but it's limited to locking down things in gconf [18:45] if it just did what we know it can without crashign, thats a very good start [18:45] sabayon was designed to let you work with almost anything [18:45] what is the purpose of tracking every change? [18:45] it has some support for Firefox and OO.o was planned [18:45] LaserJock: so if I want to set a specific bookmark for all users with sabayon, you think I'll be able to do it somewhat easily? [18:46] alkisg: yes, if sabayon is working [18:46] :-) [18:46] nubae: no, i meant external dns servers. i'm thinking about going with OpenDNS [18:46] but i can add that in the file easlily [18:46] the purpose of tracking everything is to create a complete "profile" [18:46] oh Ahmuck your server is set to 192.168.0.1 in there [18:46] how does it even work? [18:46] with sabayon basically you create a profile of what you want the desktop to look like [18:46] its set as 0.254 when u do ifconfig [18:47] then when the user logs in basically they're /home dir is replaced with the "profile" [18:47] inet --> router --> server --> switch --> client [18:47] if thats your live dhcpd.conf file, it shouldnt work at all [18:47] works great on regualar thin clients [18:47] in my script, and on the server, its set to 192.168.0.254 [18:47] like i said, if it can break, i can break it and if it shouldn't work, it works :) [18:47] yet in the dhcpd.conf file its set to 192.168.0.1 [18:48] ah, so i need to change it [18:48] * nubae scratches his head [18:48] or ur saying the fat client shouldn't even work? [18:48] right, but it shouldnt work at all [18:48] nubae, that's not the ltsp server, it's the dns server and the router, so it's working... [18:48] right [18:48] right same thing [18:48] No, in the dhcp offer the client gets server = .254 [18:49] So they boot fine, they just don't get correct dns/gateway addresses [18:49] wait.... his ifconfig shows his server has ip 0.254 [18:49] and his dhcpd.conf file shoes 0.1 [18:49] ohhh wauit [18:49] it's magic \o/ [18:49] u're right [18:49] That's what I'm saying. The dhcp server still sends server=.254 [18:49] sorry ... [18:49] ltsp-server isnt defined [18:50] If he enabled next-server, then it wouldn't boot [18:50] ok... ahmuck, change everything that is .1 to 0.254 :-) [18:50] right, thats what I expected [18:51] hmmm, and how are the thin clients getting inet then? [18:52] with a broken router and dns server [18:53] they are? [18:53] erm, that is, through the server or as fat clients? [18:54] no as thin clients they are, which is off [18:54] odd [18:54] Because if they work as thin clients, they don't need dns/gateway [18:54] oh, well I guess if its properly defined for server [18:54] They just connect to the server, and the server in turn uses it's own dns/gateway [18:54] yeap [18:54] mine dont work if I dont have proper router defined [18:54] but of course I use an external router [18:55] thin or fat? [18:55] to serve dhcp ? [18:55] both [18:55] thin should work, as long as they boot [18:55] yeah, true [18:56] yeah Ahmuck just change dhcpd.conf settings [19:01] is there a way to target a specific user with a specific fat client? [19:01] that's the next step [19:01] * Ahmuck booting fat client now [19:02] nubae: should your script have "peeked" at the default settings and copied ? [19:03] dhcpd.conf is editted manually by the admin on the server side [19:03] and the script just grabs whatever config details to connect from its side [19:03] Ahmuck: the way to set specific users is through mac:add:resses [19:05] fat client is still dead to the inet [19:06] i haven't yet sat down and played connect the dot [19:06] did u restart dhcpd3-server? [19:07] i restarted the server [19:07] before restarting the client? [19:07] yes. i did not start the client this morning until i had restarted the server [19:07] i'll restart the client again just to be sure [19:07] yes but when did u restart the server, just now? [19:08] yes [19:08] well, should actually be fine now, on the fatclient try pinging google.com [19:09] to make sure dns is working [19:09] unless u dont have a dns server running on the server, in which case, choose another address [19:10] does dns run on server by default? [19:10] iirc, it does [19:10] btw,i love edubuntu's icons [19:12] google is dead on the fat client [19:13] then its not resolving, paste me your /etc/ltsp/dhcpd.conf [19:14] Ahmuck-Jr, if I got this right, you're using 192.168.0.254 as a dns server in your dhcpd.conf, but you don't have bind installed in the ltsp server. [19:14] So you should either install bind or just change dhcpd.conf [19:14] much better explained :-) [19:14] and put whatever dns server your ltsp server has (cat /etc/resolv.conf on the server) [19:15] http://pastebin.be/15748 [19:15] alkisg: i'm using a default ltsp server [19:15] setup [19:15] k that looks fine, now do what alkisg says [19:15] by default bind is not installed [19:15] bind = dns server [19:15] outside of the modifications that nubae has had me make for the fat client setup [19:15] ah. should that be part of your fat client setup? [19:15] no :-) [19:16] bind is already installed [19:16] Also, you have a 2 nic server, right? [19:16] yes [19:16] And the clients are only seeing your first nic [19:16] oh right gateway from one to the other [19:16] inet incoming on one, and thin clients on another [19:16] yeap, masquerading [19:16] the thin clients work fine [19:16] it's the fat clients that don't work [19:17] on the inet side. the fat clients boot fine, but no inet [19:17] right, imagine u had computers elsewhere on your nework though [19:17] fatclient is the same [19:17] You need gateway & dns on the fat clients (= put them in dhcpd.conf), and to enable ip forwarding on the server [19:17] ok. i wondered about that [19:17] myself [19:18] all this shoudl normally be already set up on standard 2 nic server [19:18] An external dns server? [19:18] any dns server [19:18] I don't think dhcp3.conf gets any dns server [19:18] it gets 0.1 as default [19:19] But bind isn't installed by default [19:19] yeah I know... but what I'm saying is, this is server stuff... the fatclient script shouldnt be setting server variables [19:19] Sure, that's correct [19:20] its assumed that u have either installed bind, or u use another dns server [19:20] But Ahmuck-Jr still has to fix dhcpd.conf [19:20] yeah [19:20] but he says he has bind installed [19:21] So Ahmuck change this: option domain-name-servers 192.168.0.254 to whatever dns your server uses (cat /etc/resolv.conf on the server) [19:21] And you'll also need to enable ip forwarding on the server [19:21] one shouldn't really need that, I think enabling gateway address for the internal nic should work [19:22] I don't think so, the packets won't be passed to the other nic if he doesn't enable ip forwarding [19:23] (for the fat clients again, the thin should work without any of these stuff) [19:23] well, there is still a slight difference between fatclient and real fatclient [19:25] but to be honest, I'v not looked at the script in a while, so your're probably right [19:25] I don't know about that, it's your field! :) [19:25] I'm talking about this: [19:25] sudo su [19:25] iptables -t nat -A POSTROUTING -j MASQUERADE [19:25] echo 1 > /proc/sys/net/ipv4/ip_forward [19:25] exit [19:25] right I know, I remember not having to do that [19:25] OK, back to pascal programming... :( [19:26] :-) [19:49] still no connection [19:50] i'll have to sit down and connect the dots so to speak to see what's going on [19:52] u masqueraded? [19:52] yes, i did get my masquarde mask on. i think it looks fashionalbe :) [19:53] yes, i did [19:53] that waht alkisg posted [19:53] i'm a little confused [19:53] https://wiki.edubuntu.org/ThinClientHowtoNAT [19:56] k, i don't recall having to do this for the thin client setup. it just worked, or sorta. i did have to set one address to dhcp iirc [19:57] right this is needed only if you are running external clients, which fatclient kinda is [19:58] actually paste me /opt/ltsp/fati386/etc/resolv.conf please [20:04] http://pastebin.be/15749 [20:05] http://pastebin.be/15750 === hfsdo_ is now known as hfsdo [20:06] http://pastebin.be/15751 [20:06] sorry for all the names, i have to walk from my office to the classroom [20:06] k its detecting your nameserver as 192.168.40.40 [20:07] er, nic names [20:07] so put that in dhcpd.conf [20:07] 192.168.40.40 is a tomato router [20:07] it's there [20:07] ah, it's there as the gateway [20:07] right [20:07] so set it as dns server in your dhcpd.conf and u should see it working [20:08] so delete the gateway and set it to be the dns server [20:08] no no [20:08] just in dhcpd.conf [20:08] edit the current option domain-name-servers 192.168.0.254; [20:08] ok [20:08] sudo start dhcp3 [20:08] and then restart dhcp3-server [20:09] Ahmuck-Jr: your interfaces file is wrong [20:09]     gateway 192.168.40.40 [20:09] The gateway can't be on a different subnet [20:09] So you should delete this line, eth0 will be used when a gateway is needed [20:10] yeah gateway will be automatic [20:18] nope [20:19] Ahmuck-Jr: can you ping eth0 from a fat client? [20:25] i can ping the incoming nick as well as the thin client nic [20:26] why would i see an eth38 ? [20:27] i cannot however ping 192.168.40.40 from the client [20:27] which is the router [20:27] So you ip forwarding doesn't really work [20:29] Where do you see an eth38? [20:30] ifconfig [20:30] on the server? [20:30] http://pastebin.be/15752 [20:30] cat /etc/udev/rules.d/70-persistent-net.rules [20:31] and pastebin it [20:31] http://pastebin.be/15753 [20:32] Eeeek!!! nubae, does this have anything to do with the fat client setup? [20:32] This file should only have your 2 nics.. [20:33] yes, well i was taken aback when i saw the file as i thought it was odd as well [20:33] i do have virtual box installed here, but iirc, it's self contained [20:34] Yes, vbox doesn't have anything to do with this file [20:36] no [20:36] fatclient doesnt touch that [20:36] So, Ahmuck, better start with this one! :) Leave only your two interfaces!!! [20:36] reboot is needed, though [20:37] where are all those interfaces coming from? [20:37] And you could rename eth38 to eth1, it would make much more sense [20:37] eth0 [20:37] his eth1 is the thin client one [20:37] sorry, eth0 [20:37] Ahmuck: did you try 38 different NICs on the server?!!! [20:37] that is totally weird man [20:37] :-) [20:38] no, both nics are internal [20:38] ifconfig is handing out only 2 it seems [20:38] i thought i noticed this ealier with a firewall setup. could it be a os bug? [20:38] There are different mac addresses there, so i suppose they are real NICs [20:38] weird weird [20:38] well vbox could do that [20:38] Nah, I got vbox for a year now, never seen it touching that file [20:39] yeah but under different configs dont u get bridges instead of nat? [20:39] I've only seen multiple nics in that file when replacing nics [20:39] and a bridge would create other nics, no? [20:39] I've used it with bridges as well [20:39] k... just a guess :-) [20:39] ...and these aren't bridges!!! [20:40] ahmuck, u're a good tester... u really do find all user case scenarios ;-) [20:40] Ahmuck: did you run any mac-address changing utility? [20:41] yeah those are all real physical devices... [20:41] ...and all forcedeth [20:41] what is forcedeth? [20:41] so I guess something was changing his NIC mac address [20:42] I guess it's a vendor / product name, like realtek or something [20:42] must be abug [20:42] (nvidia google tells me) [20:42] https://bugs.launchpad.net/debian/+bug/153727 [20:42] Ubuntu bug 153727 in debian "Ethernet device's number increases by one after every reboot" [Undecided,Confirmed] [20:43] Nah.. Ahmuck's case is random, not increasing! :P [20:43] its got from 1 to 38 [20:43] that seems increased [20:43] I guess in the next reboot it'll be eth39! [20:43] hehe [20:44] Oh, I read "mac-address" instead of "device number", my fault [20:44] And forcedeth again, eh? Good catch, nubae... [20:45] so nvidia driver related it seems [20:46] Hmm... did you see the proposed workaround? It turns out that you may declare a NIC without declaring its mac address... Very useful for cloning! [20:46] hmmm [20:47] well, seems a fix would be replacing kernel driver [20:48] actually intrepid should have fixed it [20:48] I'm talking about this one: https://bugs.launchpad.net/debian/+bug/153727/comments/7 [20:48] Ubuntu bug 153727 in debian "Ethernet device's number increases by one after every reboot" [Undecided,Confirmed] [20:48] It's useful for cloning [20:48] yeah [20:50] so its a bug in forcedeth.c [20:51] Ahmuck? did you do the workaround? [20:53] wow, what an annoying bug [20:53] (A difference in Ahmuck's file as opposed to mine, is that Ahmuck's gets generated by /lib/udev/write_net_rules, while mine gets generated by /etc/udev/rules.d/70-persistent-net.rules)!!! [20:54] So there must be something else wrong also [20:55] (it says so in the header of the files, Ahmuck's and mine) [20:55] maybe something to with an upgrade vs install [20:56] I've got the same as ahmuck actually [20:56] ;-) [20:56] # This file was automatically generated by the /lib/udev/write_net_rules [20:56] # program run by the persistent-net-generator.rules rules file. [20:56] Oh, ok then [20:56] I'm on a clean installation [20:57] well, I upgraded [20:57] well, sounds like you're right! :P [20:57] but from alpha 5 Ithink [20:58] anyway its presistent-net-generator.rules creating it [20:59] anyway its definitely because of the forcedeth driver [21:00] OK, where is Ahmuck? !!! :)\ [21:13] LaserJock: if u are around... many of the apps mentioned in the spreadsheet don't have a counterpart on appnr.com [21:14] and many more educational apps are in appnr.com but not on the list... === Michiel is now known as Guest81991 === Guest81991 is now known as Eghie_ [21:18] For now I'm going to include only those that are on both, otherwise we cant take advantage of click and install [21:19] nubae: well, we need to categorize [21:20] whatcha mean? [21:20] well, we have different levels of "support" [21:21] ok, but shouldnt everything thats in edubuntu be in appnr? [21:21] since appnr has universe and multiverse === joerg_ is now known as joerg [21:22] no [21:23] Edubuntu does not come from Universe or Multiverse [21:23] it comes from Main [21:23] right, thats what I mean [21:23] and appnr.com is just a list some people made [21:23] hmmm.... I thought it took the packages as a base [21:23] all the pacakges, sorry [21:25] its a bit of a pain, because the lovely detailed explanations on appnr, for individual packages is a blessing [21:26] taking that into account actually makes sense to put edubuntu into various categories, ie. -supported, -unsupported, maybe... [21:26] well, the explanations are just from the package themselves [21:27] I guess I should be installing the packages just to make sure they work and all [21:27] nubae: did you get the hardy and intrepid lists done? [21:28] I believe I have infront of me the intrepid list, but really it should just be one list with 3 categories (hardy, jaunty, intrepid) [21:28] I think we want 1 list for each release [21:28] so something like edubuntu.org/8.10/apps [21:28] yeah, but the spreadsheet to base it off [21:29] or the whatever to base it off, a spreadsheet is a pain [21:29] a db would be best [21:29] sure [21:29] but for now I'm putting everything in an html file [21:29] the spreadsheet was just to put something down [21:29] so I can upload it to edubuntu.org as is [21:30] ok, well from what you're saying, I still need to figure out what is in hardy and intrepid, and make 2 lists for that, and then we need to decide what goes in Jaunty [21:30] right now I have one semi-finished nicely laid out html file [21:30] yep [21:30] can you put it up somewhere for me to look at? [21:31] yep, I'll put it up on a hidden node at edubuntu.org, let me just finish off what I'm doing... [21:32] np [21:33] you're sure all these apps are actually packaged, because there are some I can't find using apt-cache search or show [21:33] and I have every repo enabled [21:34] well [21:34] I think they're packaged, but the spreadsheet just has the Add/Remove name [21:34] should still find it with apt-cache show [21:34] search I mean [21:35] only if the Add/Remove name is similar [21:35] it may or may not [21:35] well it looks in description too, no? [21:35] right, but Add/Remove can be anything [21:36] ok, yeah I'm seeing... sometimes no relation [21:39] for example: CoqIde Proof Assistant does not exist [21:39] in add/remove [22:06] ok, i'm back [22:07] sorry, i have students for the rest of the day [22:07] i'll have to do this again tomorrow [22:07] no worries,.. [22:07] u saw we found the bug [22:07] its with your network card driver (forcedeth) [22:24] LaserJock: I'm gonna have to go through all the pacakge lists? ie, edubuntu-legacy, edubuntu-yound, etc, to figure out whats in each distro [22:25] nubae: use edubuntu-desktop [22:25] I'm doing it through the bzr meta package [22:25] ok [22:26] so you want the desktop-addon seed [22:26] that doesnt include everything at all [22:26] maybe ship-addon [22:27] oh wait not shipped anymore is it [22:27] k, so how does one fix that bug? [22:27] i'll look at the link [22:28] ship-addon has got a bit more but desktop-addon should have what you want [22:28] no, maybe 10% [22:28] 10% of what is listed in the spreadsheet [22:28] if that [22:29] childsplay isnt even listed there, looks like just the officially canonical supported apps [22:29] ok, but the spreadsheet is *way* more than what Edubuntu supports [22:29] childsplay is not part of Edubuntu [22:30] ok, now I'm very confused... I am using the spreadsheet as base for what is or is not edubuntu [22:30] the spreadsheet has everything that's remotely educational that in Add/Remove [22:30] far from everything [22:31] it must be based on hardy or even gutsy [22:31] shouldn't be [22:31] I think hardy or intrepid [22:31] many apps are not in add/remove, and many many in add/remove are not in the list [22:31] only conclusion I can come to [22:31] many apps are not in Add/Remove [22:32] just looking at A, we have anki, adun, avogadro [22:32] all listed in add/remove [22:32] but not in the spreadsheet [22:33] and its like that for every letter [22:33] right [22:33] I *think* the spreadsheet is from Hardy [22:33] I put avogadro in Intrepid so that would make sense [22:34] ok, so then I can conlcude this, whatver I don't find in the add/remove, has been removed since hardy, right? [22:34] hmm, not necessarily I don't think [22:34] it could be removed [22:35] well, ok, I'll just put a little N next to the ones that are listed in the spreadsheet but not in add/remove [22:35] and I'll use the add/remove in intrepid for the new base, sounds ok? [22:36] where is the spreadsheet? [22:36] but then again, we should really define what edubuntu is, cause if this is not edubuntu, then is this really what I should be doing [22:36] nubae: honestly, I wouldn't use add/remove at all [22:36] * nubae runs around in circles [22:36] I would just use the seeds [22:37] the seeds have practically nothing in them [22:37] yep [22:37] but that's what we have [22:38] you should have a list of ~20-30 [22:38] :-) so why did I get that list then :-/ [22:38] I've just been putting in hours of work with that [22:39] that's why I've been poking you about it ;-) [22:39] remember, I said to use the seeds [22:39] I understood work on the seeds for Jaunty, and for what to include in it... [22:40] gah, k, never mind... this can be the prospective edu apps list [22:40] I can pull whatever from it and stick into the actual lists [22:41] nubae: yeah, don't scrap what you have for sure [22:42] * nubae now understands what LaserJock means with just main [22:44] brackets are recommends, right? [22:44] yep [22:44] its a good thing I didnt start with the games... [22:45] games app list is humungous [22:45] hehe, yes [22:45] humungous and not in Edubuntu [22:46] but its confusing for folks if its available under Education in the add/remove menu [22:47] I mean, if I got this confused... I'm surely not the only one [22:47] I bet most people will assume education in add/remove is edubuntu [22:49] well, there certainly is some confusion I guess [22:51] that's why I think there may some value in these bundles [22:52] and separating "Edubuntu" from "Ubuntu in Education" [22:52] traditionally Edubuntu was defined by the CDs [22:53] well its easy enough, we can have 2 categories, supported and unsupported [22:54] right, but really that's not trivial [22:55] well, it requires creating another seed right?, but that can be as big or small as one pleases [22:55] maybe not if its supposed to be all encompassing though [22:55] but if you put something in a seed you "own" it [22:57] u mean u're responsible for its packaging?? [22:57] traditionally yes [22:57] for it's bugs, etc. [22:57] omg [22:57] but thats where the -unsupported comes in [22:58] right, but -unsupported still needs to be supported ;-) [22:58] software doesn't just "exist", it need to get maintained [22:58] ok, so u cant get around ignoring bugs on launchpad about something [22:59] in if you point people towards a package there is some assumption that you're going to take care of it [22:59] and sending it all downstream to the package maintainer [22:59] nah... this is like recommended stuff [23:00] but I can see this is not easy... it becomes extremely political [23:00] ok, ... but still [23:00] if we say, "go install childsplay" and then don't do *anything* towards childsplay I think it's a bad thing [23:01] u mean, just showing people the app exists, isn't enough? [23:01] I'm thinking from the users point of view now [23:02] I think they assume the software is being taken cared of [23:02] but it is, by the package maintainer [23:02] wouldn't you? [23:02] we *are* the package maintainer [23:02] that's my point [23:02] ok, I get that traditionally its that way... [23:02] but the tux 4 kids suite for example [23:03] which IS supported, all its bugs get sent downstream [23:03] not exactly [23:04] as I understand it, edubuntu maintains the seeds only, not the packages themselves [23:04] not at all [23:04] we maintain everything in the seeds [23:04] hmmm... so basically edubuntu means... supported by canonical then [23:04] no [23:05] * LaserJock thinks nubae's head is going to explode :-) [23:05] users want software that intalls and works [23:06] if it installs and doesn't work is or is buggy, that's bad for any distro [23:06] yeah, but there is such a thing as unsupported, in fact most of ubuntu apps are unsupported [23:06] no, that's a bad term [23:07] "supported" mean you can by Canonical support and that the software is maintained by Core Developers [23:08] one of the biggest advantages edubuntu has, being the 'educational distro' is introducing users to great edu apps for whatever niche they need... [23:08] and should receive better security support [23:09] "unsupported" means you can't by Canonical support and the software is maintained by MOTU [23:09] and gets security support on best-effort basis [23:10] soo all the apps in the seeds are supported officially then, right? or am I getting that wrong too? [23:11] you're right [23:13] would not that indicate that especially u should be being paid for that? [23:13] hmm? [23:14] I mean if u are the maintainer the minute u add the package to the seed [23:14] so if someone bought canonical support [23:14] u'd be paid... [23:14] umm, no [23:14] I wish it worked that way ;-) [23:15] it logically should [23:15] Canonical gets paid for Canonical support [23:15] otherwise I see no reason why one has to include only officically supported packages [23:15] right... so by including only the official packages, that only helps them [23:15] u see what I'm getting at? [23:15] hmm [23:16] no, it helps everybody [23:16] it would help more people if we included unspported pacakges too [23:17] but we can't support them [23:17] we dont need to support them, I cant believe for a moment that all kubuntu packages are supported [23:17] sure they are [23:18] I bet there are kubuntu packages that are not supported officially by canonical [23:18] not in the seeds [23:18] if it's in the Kubuntu seeds it's officially supported [23:18] right so basically edubuntu IS the officially supported educational ubuntu [23:18] right [23:19] or at least that's how it's traditionally been [23:19] well, that shows sugar out the window... [23:19] shows/throws [23:20] well, that's why we gotta work on it :-) [23:21] and who chooses what is officially supported by canonical then? [23:21] well, one has to be a bit careful here [23:21] u know what I think is maybe confusing it for me, is that within edubuntu, there is no one officially working for canonical [23:22] there used to be though [23:22] kubuntu there is [23:22] but yeah [23:22] yes, but not now [23:22] but that doesn't so much change Canonical support directly [23:22] that was of course a benefit [23:23] what we need are developers [23:23] well it does in a way... its like canonical pulling out of the educational part... or at least edubuntu [23:23] whether they come from Canonical or not is somewhat irrelevant [23:23] well, in a lot of ways they have [23:23] yeah and who wants to get caught up in it if its 'only canonically officially supported'? [23:24] I just dont see the attraction to a dev [23:24] to work on software we love [23:24] yeah, that we dont get to choose??? [23:24] I like helping people and I think education is important [23:24] sure we do [23:24] me too [23:24] incredibly important... [23:25] but we should be choosing what edubuntu is and whats in it [23:25] the problem is that there are several, sometimes intertwined layers here when you talk about "support" [23:25] we can [23:26] we can't tell Canonical what to do, i.e. what it will or won't support [23:26] but that's not everything, in fact it's fairly small [23:27] you're right, but we should be able to choose what we package/recommend [23:27] the 2 being different [23:27] we do determine that [23:27] any Core Developer can change the seeds [23:27] becoming the owner of the package... thats not a solution [23:27] and anybody can file a Main Inclusion Report to move a package from Universe to Main [23:27] well, we do that anyway [23:28] that's what being a Core Developer is all about [23:28] maintaining packages [23:28] yeah, but here we are talking about a small amount of devs and many many useful packages [23:28] yep [23:29] so we're spread pretty darn thin [23:29] the user in the end suffers, because of rigid politics that shouldn't really apply to edubuntu [23:29] thats my opinion anyway [23:29] it's not politics! [23:29] ok facts then :-) [23:29] the fact of the matter is that software doesn't just magically work on itself [23:29] somebody has got to do it [23:29] some software is better than others [23:30] look, there is no drive for a current dev to put in any more packages in edubuntu, because they cannot maintain it... that makes total sense... so edubuntu will not grow [23:30] sure there's drive [23:30] nah, if u cant, dont have the time, u can't [23:30] I want to deliver more and better educational content to kids/schools/parents [23:30] so we build a community of people [23:31] that's how this works, more people = more software maintained [23:31] if people want Universe software then it's certainly out there for them, and the MOTU do a pretty decent job of maintaining it [23:32] so the process right now would be... become a package manager, then support the app of your choice, and then suggest it for inclusion into main, before it can even start being edubuntu [23:32] that just seems like a really long, impossible process [23:32] no, no [23:32] *anybody* can contribute [23:32] I get that... but I'm talking about the apps [23:32] sure [23:32] there's no reason people can't help out [23:32] why work on an app, if it might not get into edubuntu, or there is such a long procedure for it to happen [23:33] well, because it's better for the users? [23:33] its better for the users to get an unmaintained app, than non at all [23:33] it's really not all that hard to get something into Edubuntu [23:33] not all apps, but some for sure [23:33] sure sure [23:33] childsplay is a good example [23:33] but it's in Universe, where it is maintained by MOTU [23:34] yeah, but not in edubuntu [23:34] perhaps we need a recomended list [23:34] not currently no [23:34] and default install, anything beyond that is up to the user to find out what's wrong [23:34] this is the part I'm still confused with.... there are many really good apps in uni, which I see no reason shouldn't be listed for edubuntu [23:34] nubae: then lets get them in! [23:35] how [23:35] how do we get them in? [23:35] do they all have to be free? [23:35] we file a Main Inclusion Report [23:35] yes [23:35] so for non-free apps we need a sepearte "recomended but not free" apps list [23:35] we need to do a security review and file out "paperwork" [23:35] ok, maybe getting into main is easier than I think... [23:35] ok, but hang on a sec [23:35] the security review is to prevent someone from inserting a rouge app [23:35] but I was led to assume it isnt [23:36] the problem I see is everybody is wanting to make big lists of software [23:36] when we can't even maintain the small list we have now [23:36] but many apps in uni are better maintained than these because there is MOTU [23:36] I think it's wise to start small, and work our way out [23:36] I agree [23:36] nubae: right, and that should *not* be the case [23:37] our apps should be better maintained [23:37] we had thin client manager listed for 2 releases even though its totally unsupported and dead [23:37] in its place I'd rather see a uni maintained edu app [23:37] it's not dead, it's just been replaced [23:37] dead to me = not workign [23:37] which it didnt [23:37] it works in jaunty I believe [23:37] ogra fixed up some stuff [23:38] in any case, that was an "in house" project that kinda died [23:38] well still... dead for 2 releases [23:38] but when people take off that kind of thing happens [23:38] its just an example... something that was in Uni wouldnt end up like that [23:38] right [23:38] sure it would [23:38] for thin-client-manager it would have been just the same in Universe [23:38] it might, but then next release, u pull it [23:39] but a top priority needs to be building developer community [23:39] yeah which is easier if there are more apps in edubuntu [23:39] we gotta have people working on the software [23:39] sounds like a chicken and egg situation [23:39] might i recommend only putting in stable for edubuntu [23:39] somewhat yes [23:40] Ahmuck: how do you mean? [23:40] but the difference here is that both devs and users loose out... edubuntu remains small, a shadow of what it could be [23:40] well, a stable app is one that works [23:40] Ahmuck: that is the idea right now, edubuntu has only apps which are in the main branch [23:41] Ahmuck: that is exactly what Main is supposed to be, well-working software that is maintained, stable, secure, and relatively bug free [23:41] but the problem is that its too conservative an approach, there are not enough apps [23:41] more apps doesn't really help us though [23:41] but in reality its not like that for edubuntu due to the small dev community [23:41] it helps growth [23:41] how? [23:41] more people use it, more likely to get dev interested [23:42] people aren't all that likely to use it if it's poorly maintained [23:42] otherwise the way would be.... I like app such and such [23:42] in order to work on it officialy, it needs to get into main first [23:42] thats just backwards [23:42] that's not how it is [23:42] *anybody* can work on *any* package [23:43] right they would not be working on an edubuntu package [23:43] the issue is if the work needs sponsorship or not [23:43] ok, but I don't seen anybody out there [23:43] it's just me as far as I can tell [23:43] well, and stgraber on LTSP [23:44] right... which is why I'm saying why not focus on the stuff u have on your plate, and get the MOTUS to help with other apps [23:44] and in the meantime, build the community [23:44] because MOTU doesn't have time for it [23:44] but yeah [23:45] in a while (not sure how long exactly) Main and Universe won't exactly exist, which will probably get you closer to what you want [23:45] but I'm not sure how to build a community that people don't want to participate in [23:46] look, why dont we at least look at the edu apps that are currently in universe, and see what would be cool for inclusion, like a recommended list... we dont even have to put it in the seeds [23:46] who said they dont wanna participate, I wanna particicpate [23:46] sure, no problem [23:46] well, to be frank, education is the worst area I've seen for getting people to actually do anything [23:46] people just don't step up very often or very much [23:47] and I totally understand that [23:47] and I know at least another 5 other people... but I think not everyone can make major commitments like becoming package manager of every app they recommend [23:47] but it's a real struggle [23:47] no, no [23:47] you don't have to [23:47] we work as a team [23:47] you do what you can, when you can [23:47] ok, well we still have that strategy doc to look at too :-) [23:47] but people need to help [23:48] yeah I get that, maybe a list should be made [23:48] ask what people want to work on [23:48] I need to figure out how to make Edubuntu appealing to 1) current devs and 2) educators who can become devs [23:49] well there needs to be a procedure too [23:49] its not that simple to get involved right now... well... not simple enough [23:49] right, it's not terribly simple [23:49] I'm not sure if I can do a whole lot about that :( [23:50] well what do u tell someone when they ask what shall/can I do to help with edubuntu [23:50] one thing that I've thought about is getting an MOTU Education team [23:50] and most of these will say, I have little programming experience [23:51] right [23:51] the problem right now is we don't exactly have a lot of useful infrastructure for people to plug in to [23:51] back in the day it was a heck of a lot easier [23:52] right now I'd say the big things are 1) bugs 2) documentation 3) website [23:52] well I've heard it said several times that installing edubuntu-addon was pointless [23:52] because it didn't really include a functional educational desktop [23:52] ok [23:53] we gotta take very little steps though [23:53] starting from where we are at right now [23:53] yeah I see that [23:54] so what kind of things can I/we do to help people get involved? [23:55] well next meeting get a kinda roll call and find out who would want to be involved with what [23:55] and list it somewhere [23:56] maybe make it easy to do the same via a wiki/online form on the website [23:57] or even form to email [23:57] should there be a list of tasks that need somebody to work on them? [23:57] yeah that too [23:57] and just state, this is not an exhaustive list [23:58] normally that would be on https://wiki.ubuntu.com/Edubuntu/Devel/RoadMap [23:58] i think that link would scare people away [23:58] especially with the work devel in it :-) [23:58] s/work/word [23:58] hmmpf [23:58] it shouldn't [23:59] reaction might be, I'm not a developer, thats not for me [23:59] but I guess it could, it's somewhat hard for me to think like a newb anymore :( [23:59] right [23:59] very few people like to call themselves devs [23:59] or dare [23:59] hmm [23:59] in Ubuntu we use if very loosely