20:59:21 #startmeeting 20:59:21 Meeting started Mon Aug 30 20:59:21 2010 UTC. The chair is brunowolff. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:59:21 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:59:30 #meetingname spins-sig 20:59:30 The meeting name has been set to 'spins-sig' 20:59:31 * nirik is around. 20:59:43 #topic Who's here 20:59:53 #chair nirik 20:59:53 Current chairs: brunowolff nirik 21:00:34 #topic compose status 21:00:44 everything is composing currently... 21:00:48 including games. ;) 21:00:59 Too give a bit of time for people to show up let's do the FYI stuff fist. 21:01:12 ok. 21:01:16 Games is real close to the limit though. 21:01:26 many spins need to work on size limits. 21:01:33 Perl is still getting pulled in. 21:01:34 * nirik is going to look at xfce for that soon. 21:01:58 yeah, anaconda->vnc I think 21:02:00 Once there is a new F14 anaconda that should get fixed. The other stuff is in place. 21:02:25 livecd-desktop is over size. Not pulling in perl will make it very close. 21:02:32 https://bugzilla.redhat.com/show_bug.cgi?id=627280 21:02:56 I think the new tigervnc is in place. For sure its either in stable or testing. 21:03:25 #info spins composing, need to watch sizes for some spins. 21:03:33 finishing the gtk2 downbuilds would help too. 21:04:02 #topic FYI livecd-tools 21:04:34 I want to try to get a new livecd-tools out this week. That's about as late as I'd want to go with other than bug fixes. 21:04:52 There are a couple of interesting scripts being added, that shouldn't break anything. 21:05:09 ok. 21:05:14 A couple of fixes, and maybe a change to --skipcopy. 21:05:54 Jasper has been doing a lot of work with this, but I need to go through the list for patches making sure stuff is upstream for the new build. 21:06:13 If something bad happens we can go back to 033. 21:06:40 #topic fedora repo changing in git repo 21:06:55 #link https://fedoraproject.org/wiki/Spins_repo_directives 21:07:27 I found the few places that supply repo directives aren't very consistent. 21:07:54 I think the install ks is an exception and don't want to make it the same as the others. 21:08:03 agreed. 21:08:10 So I'd like to make the other three the same. 21:08:13 we should make all the other 3 the same. 21:08:14 yeah. 21:08:34 I'd like them to not have release numbers, but rather use the releasever macro. 21:08:51 ok, what populates that? 21:09:01 The timing would be that when Fedora branches we change the master branch to use fedora and fedora-updates. 21:09:34 And then at the release (after branching the repo) we change master back to using the rawhide repo again. 21:09:35 ok, although sometimes when branching there is no updates repo, just updates-testing. 21:09:55 sounds good to me otherwise. 21:10:03 I think the mirror manager is supposed to handle that. 21:10:12 yeah, could be 21:10:22 It should point to an empty repo. 21:10:41 yum does the macro substitution. 21:10:57 I think that is a relatively new feature. 21:11:07 ? 21:11:10 releasever? 21:11:12 it is not new 21:11:18 There is also a command line override, but no way to pass one through livecd-creator. 21:11:20 brunowolff: how do we set that with livecd-creator I mean 21:11:55 livecd-creator will use whatever version it finds. 21:11:59 so it defaults to whatever the current machine's value is. ? 21:12:22 In mock it would pick up the version from fedora-release in the chroot. 21:12:58 It may also look at environment. 21:13:28 For most testers, I suspect they'll be doing custom stuff to point to their own local repos. 21:13:41 ok 21:13:53 The nightly compose script seems to do something special to use the correct mirrors as well. 21:14:27 yes, it seds the lines to the correct value 21:15:09 I didn't see much in the yum man page other than for the command line option. 21:15:29 Do you think that making this change now will break the nightly composes? 21:15:47 it shouldn't... and it should be easy to fix if it does. 21:16:19 OK, then I think we'll go ahead with this. We have time to tweak things if there are issues. 21:16:44 #action brunowolff will implement https://fedoraproject.org/wiki/Spins_repo_directives 21:17:02 #topic odd install ks repo directives/ 21:17:37 I noticed the install ks has --excludes on its repo directives. 21:18:07 I am guessing is that this blocks things harder than using -packagename in the %package section? 21:18:30 Do we have any problems with releng doing this in the install script? 21:18:50 If someone wanted to do this for a live image would we have problems with it? 21:19:11 I think that install is releng's domain and don't have a problem with that. 21:19:14 yeah, seems odd, but there's likely a reason. 21:19:20 we could ask them about it for our info... 21:20:06 I think it would be a bad idea for live images, as doing things like --excludes=perl would remove anaconda (currently) which is a no no. 21:20:40 There are also some odd things like btanks. 21:20:57 yeah. 21:21:24 Does Jesse usually do that? 21:21:54 Or are their multiple people likely to mess with that file? 21:22:08 I suppose I could do a git blame or something to see. 21:22:14 I think it's only possibly notting or Oxf13 21:23:16 #action brunowolff will ask releng about the install ks as an FMI. 21:23:56 We don't need to decide about live images doing that right now as no one is doing it. 21:24:20 #topic others 21:24:38 an FMI ? 21:24:46 I only got the one document done, so I don't have testing recommendations yet to review. 21:24:54 For My Information. 21:25:06 what was the question, maybe I can help now? 21:25:17 I wanted to ask why you guys use --excludes instead of -packagename? 21:25:33 because pungi is an inclusive gatherer 21:25:43 it will gather all matches of a particular package name or requirement 21:25:48 I figured it was to make sure the package was blocked and not brought in when dep solving. 21:26:05 and the way that kickstart semantics are things that are -foo can still be brought back in via requirements 21:26:18 so if we really don't want a particular package in our composition, we have to exclude it from the repo itself. 21:26:28 OK, that's what I guessed. 21:27:05 Does that get reviewed from time to time? I saw btanks was excluded which seemed odd to me. 21:27:19 not often 21:27:23 git blame would help 21:28:05 Thanks for the answers. That will help if we get a request for someone to do this for a live image. 21:29:38 Does anyone have anything else for today's meeting? 21:30:35 #action brunowolff will send out meeting notes. 21:30:53 That sounds like a wrap for today. Thanks for attending! 21:30:59 #endmeeting