21:13:20 <biertie> #startmeeting spins-sig
21:13:21 <zodbot> Meeting started Mon Feb  8 21:13:20 2010 UTC.  The chair is biertie. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:13:22 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
21:13:33 <biertie> #topic agenda
21:13:41 <biertie> status games spin
21:13:48 <biertie> status security spin (?)
21:13:53 <biertie> status mini kickstart file
21:14:01 <biertie> and open floor
21:14:15 * nirik can comment on several issues with nightly composes if people care.
21:14:32 <cwickert> go ahead nirik
21:14:34 <biertie> and just before the open floor: issues with nightly buils
21:14:41 <biertie> #topic games spin
21:15:01 <brunowolff> The games spin is still oversized, but I will be cutting stuff
21:15:12 <brunowolff> before the devel freeze to get it under one way or the other.
21:15:36 <brunowolff> The QA guys made some cuts in their spin, and I'll look at what they did.
21:15:50 <brunowolff> Hopefully the desktop spin won't get much larger before release.
21:15:54 * sdziallas will probably cut some stuff from soas & the design-suite, too.
21:16:16 <brunowolff> glest is still not building after the xerces-c update, but I own that package now and think I can
21:16:28 <brunowolff> get that fixed before the devel freeze as well.
21:16:48 <brunowolff> That's it.
21:16:58 <biertie> so next topic? :)
21:17:22 <brunowolff> Yes, move on.
21:17:26 * huff here
21:17:33 <biertie> #topic security spin
21:17:51 <biertie> but I'm not sure we can talk about that, I again forgot when the board meets..
21:18:03 <biertie> I have to do a brain transplant, I think
21:18:35 <cwickert> the spin is ready for first testing, but not yet finished
21:18:59 <biertie> moblin spin is also still in 'ready for board' cat
21:19:06 <biertie> you have a lot of legal issues sdziallas ;)
21:19:12 <sdziallas> that was approved, afaik.
21:19:20 <sdziallas> actually, that's not mine :) (but pbrobinson's)
21:19:24 <cwickert> biertie, it was approved by the board
21:19:32 <sdziallas> biertie: I try to stay clean... ;)
21:19:52 <biertie> sdziallas: oh, I'll move it myself then :)
21:20:10 <cwickert> let's do the security spin forst, one after the other
21:20:28 <cwickert> so, it's ready for basic testing
21:20:34 <cwickert> but it needs artwork
21:20:41 <cwickert> and an additional polkit package
21:20:49 <cwickert> for now, we use sudo
21:20:54 <cwickert> not nice, but it works
21:21:07 <cwickert> but we will switch to polkit soon
21:21:29 <cwickert> and we will also switch to another terminal, most likely roxterm or Xfce's terminal
21:21:49 <cwickert> so what you get with the alpha will not be the final result but it's ok for testing
21:21:50 <cwickert> eof
21:24:27 <biertie> hehe, ok :)
21:24:46 <biertie> something else on security spin?
21:25:00 <cwickert> any questions?
21:25:29 <biertie> not for me :)
21:26:38 <biertie> next topic then :)
21:26:51 <biertie> #topic kickstart builds
21:27:12 <biertie> wrong topic, but this order will work too, I guess :)
21:27:21 <biertie> nirik: ping
21:27:33 <nirik> yes, where are we?
21:27:59 <nirik> the info about nightlys?
21:28:02 <biertie> ach
21:28:09 <biertie> #topic nightly builds
21:28:16 <biertie> that topic was a bit too wrong ;)
21:28:25 <nirik> so, a number of the spins were failing over the last week or so.
21:28:28 <biertie> anyway, info about the nightlys, yes :)
21:28:32 <nirik> This was due to two big issues:
21:28:53 <nirik> 1. There were broken deps in ibus-* packages. This prevented a number of the spins from composing.
21:29:08 <nirik> this was fixed today, so should be ok for tomorrow.
21:29:30 <nirik> 2. There was a resize2fs issue that broke soas and the i686 design-studio spins.
21:29:42 <nirik> it should also now be fixed from what I can see.
21:29:53 <nirik> So, those should have composed ok today.
21:30:05 <nirik> https://bugzilla.redhat.com/show_bug.cgi?id=562605
21:30:07 <nirik> is the bug for that one.
21:30:23 <nirik> Hopefully tomorrow we will get a full slate of composes. ;)
21:30:40 <nirik> Also, I am going to try and fix things so it won't remove an old iso if there is no new one composed...
21:30:42 <brunowolff> Games spin probably won't work tonight.
21:30:48 <nirik> so we should have at least the last one that worked.
21:30:51 <brunowolff> It needs glest fixed.
21:30:53 <sdziallas> nirik: nice! (awesome work, btw)
21:30:54 <nirik> brunowolff: ? oh, right.
21:31:04 <nirik> thats all I had. ;)
21:33:00 <biertie> ok, any questions / things anyone wants to add, to that?
21:34:56 <biertie> ok, I'll tak that as a no :)
21:35:05 <biertie> #topic mini ks
21:35:10 <biertie> huff: ping
21:35:25 <huff> yo
21:35:31 <huff> ok
21:36:02 <huff> so ive made a alot of progress on the mini ks however im not sure if its goign to work on the moblin stuff
21:36:12 <huff> I think they are two differnt ideas
21:36:27 <huff> I am working on base fedora image no x
21:36:34 <pbrobinson> huff: where can I see what you've done?
21:36:35 <huff> I think he wants somthing else
21:36:40 <huff> pbrobinson: yea
21:37:06 <huff> Its basicly aos stuff with some abstraction
21:37:10 <pbrobinson> well the base with no X will still be something I can build on
21:37:19 * sdziallas hullos (interested, too)
21:37:39 <huff> pbrobinson: biertie: I just courious as to how ppl would consume this
21:37:55 <pbrobinson> my movement is somewhat slowed due to travel and house moving at the same time
21:38:05 <huff> ie. do we want a base ks. or just a base-package set
21:38:58 <biertie> well, if you aren't building a server thing, you need X right
21:39:15 <biertie> or, could the mini ks -without x- also work for embedded stuff then?
21:39:18 <pbrobinson> huff: ultimately I would have thought you mini.ks would replace the official one, then people can include it in their .ks so we might have mini.ks and a mini-x.ks
21:39:45 <pbrobinson> with the later being basically what mine is, but including and adding to yours
21:40:16 <huff> pbrobinson: yes that sounds good to me but i would think mini-x.ks would build on top of mini.ks
21:40:17 <pbrobinson> is yours basically core + networking, does it have yum?
21:40:24 <huff> pbrobinson: yes yum
21:40:27 <huff> dhcp
21:40:33 <pbrobinson> huff: yes, exactly!
21:40:41 <huff> selinux
21:41:07 <pbrobinson> so core to get a booted, networked box. sshd?
21:41:17 <huff> pbrobinson: yep
21:41:30 <huff> i think sshd but I have heard otherwise
21:41:40 <huff> but currently I include it
21:41:52 <cwickert> is it installable?
21:41:59 <huff> cwickert: ?
21:42:00 <cwickert> so is anaconda-runtime in?
21:42:06 <huff> no
21:42:16 <nirik> isn't that just core+base ?
21:42:18 <huff> I do not include that
21:42:38 <cwickert> IIRC one requirement for a "spin" is that it's installable
21:42:47 <huff> I use --nobase
21:43:17 <huff> currently 156 packages ~ 430M
21:43:28 <sdziallas> well, I guess the deal is to have something that we can base our stuff on. if all spins just randomly include @core and @base, that's probably not the best way (which is why I like the mini.ks idea, actually)
21:44:19 <pbrobinson> nirik: core+base are way too big!
21:44:32 <nirik> yeah, ok. I see what you're saying...
21:44:49 <nirik> so, what do we do? you guys want to work on a mini/mini-x ?
21:44:55 <pbrobinson> cwickert: well its not a spin as such. Its a building block to make a spin
21:45:08 <huff> im just having trouble picturing the bes way to consume it, ie so that others can use it w/out having to chage stuff
21:45:18 <huff> ie there is not a good way to override stuff
21:45:27 <huff> like i want most of it but not that
21:45:40 <brunowolff> Spins can include ks files. That's how the desktop spin gets used by the games spin.
21:45:42 <huff> am I making sense?
21:45:53 <brunowolff> I would expect the mini spins to get used the same way.
21:45:54 <huff> yes
21:46:06 <huff> but what about partion size
21:46:14 <huff> if thats defined in mini-ks
21:46:15 <brunowolff> You can override that.
21:46:18 <huff> ok
21:46:31 <huff> how does it know which one to use./
21:46:31 <brunowolff> The games spin uses an override now as the default is too small.
21:46:32 <huff> ?
21:47:19 <brunowolff> The top level one get's used. Probably its the last one seen.
21:47:37 <huff> ok thats what I thought however ive seen some weirdness
21:47:58 <huff> especially with ks flatten
21:48:51 <brunowolff> In the games spin the part command is after the include which is consistent with the "last one seen" theory.
21:49:11 <huff> do you guys see any issue with just using AOS.ks as fedora-mini.ks
21:49:19 <huff> its pretty small as it is
21:49:35 <huff> and we can take patches for any additions and/or deletions
21:50:00 <huff> but it does not have any livecd stuff, just produces a base Fedora image
21:50:31 <huff> after all that was its original intention
21:51:08 <pbrobinson> huff: is the latest AOS.ks in git?
21:51:14 <huff> yes
21:51:43 <cwickert> but IMO 430 M is still a little fat. LXDE spin is 465 M with X, LXDE, anaconda, and some apps
21:52:03 <huff> cwickert: doyou have yum?
21:52:17 <pbrobinson> because from the stuff that I did with the f-mini.ks there would need to be changes to comps to make it cleaner but I'm not sure how this is handled in aos.ks
21:52:25 <cwickert> huff, of course, even gnome-packagekit
21:52:40 <nirik> are there any spins just added to git that I should add to the nightly script?
21:52:52 <pbrobinson> cwickert: you talking about .iso size or installed size?
21:52:55 <huff> pbrobinson: well I am open for suggestions
21:53:03 <cwickert> oops, the iso
21:53:06 * cwickert hides
21:53:28 <nirik> desktop design-suite xfce broffice.org kde electronic-lab games lxde soas (list I have currently)
21:53:31 <pbrobinson> I think huff is referring to installed (moblin iso is around 400Meg too L:-)
21:54:27 <huff> pbrobinson: what exactly do you mean by changes to comps?
21:54:50 <huff> I would be open to looking in to this but not to famiular with comps stuff
21:55:51 <cwickert> I doubt that changes in comps are sufficient. for example firstboot -> system-config-keyboard ->metacity -> tons of gnome is in the requiements of the packages
21:56:28 <cwickert> huff, http://fedoraproject.org/wiki/How_to_use_and_edit_comps.xml_for_package_groups
21:59:21 <pbrobinson> huff: well @base in the .ks pulls in a heap of stuff. Same with @core. they should be updated so a mini.ks should be just @base (or @core - which ever one is smaller - they seem to be similar) and that should be it.
21:59:57 <huff> my understanding is you always get @core regardless
22:03:40 <biertie> huff: +1
22:05:18 * nirik notes we are out of time here.
22:06:23 <nirik> anything else before we close out to allow rel-eng in ?
22:06:28 <huff> pbrobinson: if you could give me some feedback/suggestions on the AOS ks file I think the easyest course of action would be to make it the Fedora-min.ks
22:06:39 <brunowolff> And it looked like releng wanted to meet right after us.
22:06:40 <huff> and expand on that for X-mini
22:08:47 <pbrobinson> huff: yea, but for instance @core has cronie, which requires /bin/sendmail, which by default pulls in exim and then perl!
22:09:15 <pbrobinson> huff: will take it offline
22:09:27 <huff> pbrobinson: ok
22:11:31 <biertie> oh, ok, I should stop the meeting then, I guess
22:11:34 <biertie> srry :)
22:11:42 <biertie> #endmeeting