bstinson / centos / centos.org

Forked from centos/centos.org 4 years ago
Clone
Blob Blame History Raw
16:20:21 <quaid> #startmeeting
16:20:21 <centbot> Meeting started Thu Jan 23 16:20:21 2014 UTC.  The chair is quaid. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:20:21 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:20:53 <quaid> #info around Java jars, all that matters is redistributability -- if you want to maintain N jars for your users, that's up to you
16:21:45 <quaid> #idea branding forum to discuss and resolve, cf. fedora-legal list
16:22:09 <quaid> #action quaid to talk with fontana about branding forum
16:22:20 <quaid> anyone else who wants to help keep notes from the meeting, go ahead
16:22:31 <quaid> I'm not sure where the logs write to yet ...
16:23:00 <DrBacchus> You'll get a URL when you end the meeting.
16:23:17 <DrBacchus> Several of them, in fact.
16:23:27 <goozbach> quaid: you'll need to add whoever wants to add notes to #chair
16:23:50 <quaid> right, the last one I saw from output didn't show the full URL just the local path, not sure what host it's on in other words
16:24:12 <quaid> goozbach: I think anyone can do #idea and #info, only #agreed and #topic and a few others are saved to chair iirc
16:25:12 <Evolution> we ( Bahhumbug ) will be improving centbot in the coming days, so there should be some additional improvements
16:25:20 <DrBacchus> Someone needs to mute.
16:25:27 <goozbach> There definately is a call for something around rebrandability/interoperability SIG
16:26:02 <goozbach> Evolution: I've got a patch for meetbot which sends email to a list when finished
16:26:05 <goozbach> if you're interested
16:26:18 <quaid> #info breaking out CLoud Infra and Instance as different SIGs
16:26:55 <quaid> goozbach: that sounds great, my least favorite part of it is having to construct that email at the end :)
16:27:04 <mburned> does the cloud instance sig exist yet?
16:27:12 <kbsingh> mburned: yes
16:27:12 <quaid> #info yes, oVirt is distributed via RPMs
16:27:13 <pixelb> Yes RDO uses either packstack or foreman as separate installers (puppet underneath)
16:27:14 * mburned probably needs to pay attention to that too
16:27:27 <kbsingh> http://wiki.centos.org/SpecialInterestGroup/CloudInstance
16:27:38 <mburned> kbsingh: thanks
16:27:40 <quaid> #info ROD users either packstack or foreman as separate installers (puppet underneath)
16:28:45 <hughesjr> ok, so then for the most part, people just want content installed and then configured later
16:29:28 <hughesjr> that lends itself to these SIGs just being repos and no changes to anaconda or isos
16:29:46 <pixelb> hughesjr, I concur
16:30:21 <hughesjr> that makes the delivery system much easier if we can get away with it
16:30:25 <kbsingh> yea
16:30:36 <mburned> i might argue that adding additional repos and/or pre-defined package sets might be useful
16:30:36 <kbsingh> but that would need to fit into the overall scope
16:30:41 <DrBacchus> And then variants - liveCDs, perhaps - of preconfigured systems, like the RDO one.
16:30:42 <Evolution> goozbach: yes please, or send it off as a pull request on github
16:31:18 <goozbach> for the voip sig (at least for FreePBX) we want an easier way to rebrand/re-relase isos/anaconda
16:31:37 <mburned> DrBacchus: right
16:32:09 <goozbach> I think the hardest part of getting a custom "installer" is the lack of documentation/tools on how to create a new spin
16:32:23 <kbsingh> goozbach: i disagree
16:32:47 <Bahhumbug> goozbach: Any information you can provide me regarding that patch would be welcome.
16:33:38 <quaid> ke4qqq: another trick of having your bits in a central yum repo is that, if Foo is in git.centos.org, Foo can be called CentOS - this resolves the question that gregdek had earlier, right?
16:33:54 <goozbach> kbsingh: ok well the "how" isn't too bad it's the "what" needs to changed to comply with trademark issues
16:34:07 <quaid> ke4qqq: we're looking at a future leg of the community build system that is Coprs-like - either user Coprs or build it in to Koji, for example
16:34:13 <goozbach> Bahhumbug, Evolution I'll have the patch up on github shortly
16:34:25 <Bahhumbug> goozbach: Thank you.
16:34:47 <Evolution> goozbach: yes. thank you. the more work I can pile on Bahhumbug the better
16:34:48 <Evolution> :-P
16:34:59 * Bahhumbug hides
16:35:14 <quaid> ke4qqq: meaning, we can lower the barrier to doing scratch builds from accepted source, making those repos available, etc.
16:39:46 <quaid> yeah, folks are already doing all this ship and support bundled libs, not going to be any harder to do in CentOS
16:41:35 <goozbach> Bahhumbug: Evolution the code as-is is here: https://github.com/gooseproject/meetbot/blob/master/ircmeeting/meeting.py
16:41:50 <goozbach> don't have a good pull request as it's spread across three different commits
16:42:12 <goozbach> and I don't have time to re-do it as proper branch/and --squash
16:44:46 <mikem23> Koji supports building with Maven actually
16:44:57 <mikem23> though the feature is not used the the Fedora instance
16:45:47 <Bahhumbug> goozbach: Thank you.
16:46:09 <quaid> I missed that
16:46:12 <quaid> what is the common thing therre?
16:46:23 <quaid> #chair Evolution
16:46:23 <centbot> Current chairs: Evolution quaid
16:47:00 <quaid> what was the question KB just got +1 all around for?
16:47:02 <goozbach> we'd (FreePBX) like the that pushbutton iso dealio too
16:47:03 <quaid> something about live CD?
16:47:06 <quaid> ah
16:47:24 <quaid> #agreed worth working on a pushbutton ISO tool as part of the CentOS Project
16:48:48 <mikem23> koji also supports building livecds and virt images
16:49:04 <quaid> #idea have upstream merit be the basis for who gets commit access, rather than CentOS having to track it
16:49:13 <hughesjr> mikem23: well then set it up :D
16:50:51 <quaid> #idea within CentOS Project tools, have it be possible for someone in the SIG to get tired of greenlighting patches from a known good person, so proposes that person to the SIG for commit access directly -- this would happen outside of the upstream's own contributor growth pathway
16:50:59 <quaid> jzb: is the above fair?
16:51:08 <Evolution> lets get some authentication in place, so that we can do much of this.
16:51:15 <quaid> #chair jzb ke4qqq mikem23 samkottler
16:51:15 <centbot> Current chairs: Evolution jzb ke4qqq mikem23 quaid samkottler
16:51:17 <Evolution> quite a bit of what we're dealing with depends on auth.
16:51:20 <jzb> quaid: yes
16:51:24 <quaid> #chair mburned
16:51:24 <centbot> Current chairs: Evolution jzb ke4qqq mburned mikem23 quaid samkottler
16:51:44 <kbsingh> the livecd and image from git stuff already works, so we dont need to block on other stuff. its a simple low hanging fruit thing that we can use to setup a relationship with
16:52:13 <quaid> #info the livecd and image from git stuff already works, so we dont need to block on other stuff. its a simple low hanging fruit thing that we can use to setup a relationship with
16:53:08 <quaid> ke4qqq: jzb I think we don't have consensus on how to handle merit and commit access to git.centos.org, I suspect it will be a blend of those two ideas; let's continue that discussion in The Usual Places
16:53:41 <jzb> quaid: might be something to handle organically?
16:53:55 <jzb> quaid:  as it happens, rather than trying to put all the ideas in place day one?
16:53:58 <jzb> (day one-ish)
16:54:05 <quaid> #idea Seed a few committers to git.centos.org from each upstream, add new commiters as per $formula_to_be_determined but which could be a combination of merit-within-CentOS and merit-within-$upstream
16:54:24 <quaid> jzb: +1 sure, i.e., no-new-thread-needed :)
16:54:41 <quaid> #idea handle organically, don't sweat
16:54:55 <quaid> #agreed CentOS is using the meritocracy spotlight, somehow :)
16:55:25 <goozbach> <-- doing anaconda "stuff" now
16:56:12 <quaid> sounds like an episode of "Dirty Jobs"
16:56:55 <hughesjr> :)
16:57:59 <jzb> quaid, kbsingh is there any thought to creating SIG-specific mailing lists?
16:58:16 <quaid> jzb: I think we were going to start on devel and split if needd
16:58:21 <DrBacchus> jzb: As I understood it, it'll be on centos-devel until it gets too noisy.
16:58:24 <quaid> it's been a rather quiet list
16:58:33 <quaid> kbsingh: can I help you with getting your notes in to this meetbot instance now?
16:58:36 <goozbach> quaid: it is a dirty job
16:59:02 <quaid> DrBacchus: yeah, I refer to that as "standard operating procedure" - wait until it's annoying, then split :)
16:59:05 <kbsingh> jzb: yes, once the SIG needs it and there is traffic that is SIG specific, lets mailing list it ( eg. CentOS-Virt is a list )
17:00:43 <jzb> kbsingh: gotcha
17:03:33 <quaid> kbsingh: do you have any notes for the meeting before I close it?
17:03:49 <kbsingh> quaid: just that i will write up notes from the call, and propose
17:10:54 <quaid> kbsingh: OK
17:10:58 <quaid> closing the meeting in a moment
17:13:15 <kbsingh> quaid: ta
17:17:32 <quaid> #endmeeting