14:31:14 <mboddu> #startmeeting RELENG (2017-06-05)
14:31:14 <zodbot> Meeting started Mon Jun  5 14:31:14 2017 UTC.  The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:31:14 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:31:14 <zodbot> The meeting name has been set to 'releng_(2017-06-05)'
14:31:14 <mboddu> #meetingname releng
14:31:14 <zodbot> The meeting name has been set to 'releng'
14:31:14 <mboddu> #chair dgilmore nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin
14:31:14 <zodbot> Current chairs: Kellin dgilmore masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll
14:31:14 <mboddu> #topic init process
14:31:42 * masta is around
14:31:53 <Kellin> .hello kellin
14:31:54 <zodbot> Kellin: kellin 'None' <kellin@retromud.org>
14:31:56 <contyk> .hello psabata
14:31:57 <zodbot> contyk: psabata 'Petr Šabata' <psabata@redhat.com>
14:32:12 <dustymabe> .hello dustymabe
14:32:13 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
14:32:41 <mboddu> FYI, nirik is on PTO this week
14:32:59 <masta> hello contyk, nice to see you last week in person.
14:33:22 <mboddu> masta: are you back in states or still in Brno?
14:33:39 <masta> back in Dallas, USA
14:33:52 <mboddu> masta: okay
14:33:55 <contyk> masta: ah, that's you :)
14:34:15 <masta> contyk, yes!
14:34:59 <mboddu> lets give couple more min for people to join and then we will start
14:35:14 <mboddu> sharkcz: are you around?
14:35:33 <dgilmore> hi all
14:35:36 <sharkcz> mboddu: have another meeting running now ... :-(
14:35:38 * mboddu wants to get an update on Alt arch, its been a while since we were busy on other stuff in previous meeting
14:35:52 <maxamillion> .hello maxamillion
14:35:53 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
14:35:55 <mboddu> sharkcz: okay
14:36:00 <mboddu> Lets start
14:36:03 <maxamillion> .pushduty
14:36:05 <zodbot> maxamillion: The following people are on push duty: Rob, Fedora Release Engineering
14:36:08 <zodbot> maxamillion: - https://apps.fedoraproject.org/calendar/release-engineering/
14:36:23 <mboddu> #topic #6822 using include statements in pungi-fedora
14:36:32 <mboddu> #link https://pagure.io/releng/issue/6822
14:36:57 * dustymabe waves
14:37:19 <masta> maxamillion, thanks,.... I was worried I was on push duty. the calendar seem to be different in recent months
14:37:41 <mboddu> dustymabe: Do you remember whats lsedlar recommendation on this? If I remember correctly, its not advised but very useful to use, I might be remembering it wrong
14:38:03 <dustymabe> mboddu: i think lsedlar's take was "why aren't we doing this already"
14:38:09 <maxamillion> masta: :)
14:38:15 <maxamillion> dustymabe: \o
14:38:17 <lsedlar> there's no reason not to recommend it
14:38:34 <mboddu> lsedlar, dustymabe : okay, thanks
14:38:48 <dustymabe> dgilmore, do you know of a reason why we aren't doing this today?
14:39:10 <mboddu> dgilmore: so, can we start using the "include" option for pungi configs?
14:39:15 <dgilmore> dustymabe: because we are not
14:39:25 <dgilmore> the use of them is not documented
14:39:39 <dgilmore> there is no reason why we can't
14:39:42 <masta> the include thing is fine
14:39:46 <dustymabe> dgilmore: cool - good to know
14:39:47 <lsedlar> true, the format generally lacks any documentation
14:39:48 <dgilmore> though for some thing it may get tricky
14:40:13 <dgilmore> like keeping a historical record that this was the config used for Alpha, BEta, Final etc
14:40:28 <masta> but, I say we run the pungi validation script on anything that uses %include
14:40:41 <dustymabe> dgilmore: i'm thinking we have a single config for a branch that is like basically a library of configuration
14:40:41 <dgilmore> dustymabe: there is no intentional decision either way
14:40:50 <dgilmore> dustymabe: huh
14:40:59 <dustymabe> and then individual pungi configs will change only what they need to
14:41:11 <maxamillion> lsedlar: may I request documentation be made? :)
14:41:19 <dgilmore> I do not grok what you just said dustymabe
14:41:50 <dustymabe> dgilmore: basically most configuration lives in one file - then individual configs (like atomic nightly cloud nightly, etc, just #include that one)
14:42:02 <dustymabe> and change what they need to
14:42:11 <lsedlar> maxamillion, yes you can :)
14:42:19 <masta> FTW  pungi-config-validate
14:42:20 <dgilmore> dustymabe: I think you need to use some more words
14:42:25 <dgilmore> and be a bit clearer
14:42:53 <dustymabe> does anyone else not know what I'm saying?
14:43:25 <contyk> you want to include config files in other config files
14:43:26 <dustymabe> i guess I just need to create an example
14:43:38 <dgilmore> contyk: that I get
14:43:46 <masta> dustymabe, yes, I've no idea what you say. please express
14:43:54 <dgilmore> but what he is proposing as to how he thinks it will work I do not
14:44:11 <maxamillion> lsedlar: thanks :)
14:44:12 <dustymabe> ok - basically there is a lot of config in fedora.conf
14:44:23 * dgilmore notes that the use fo include is not documented
14:44:32 <contyk> dgilmore: that might be a problem :)
14:44:36 <dustymabe> a lot of stuff at the top that doesn't actually tell pungi to build anything
14:44:54 <dustymabe> but is just configuration, like names, etc..
14:44:58 <dgilmore> and for things like alpha and beta and final we likely do not want to use includes so we have a record of exactly what was used
14:45:16 <dustymabe> dgilmore: personally, i think that's what git is for
14:45:29 <dustymabe> but i don't have a problem with that
14:45:33 <dgilmore> dustymabe: that only helps whjen you know the commit used
14:45:38 <dgilmore> dustymabe: which we do not know
14:46:04 <dustymabe> dgilmore: probably the last commit that edited the file
14:46:05 <contyk> wasn't there a proposal to build composes in koji?
14:46:16 <contyk> for exactly this reason?
14:46:29 <dgilmore> dustymabe: are you proposing that we have a snipped for each thing and include them all
14:46:32 <contyk> I don't know any specifics, just heard something like that
14:46:36 <dgilmore> so there would be a Atomic host snippet
14:46:39 <dgilmore> workstation
14:46:41 <dgilmore> Server
14:46:43 <dgilmore> Cloud
14:46:51 <dgilmore> spinad and labs etc?
14:46:57 <dustymabe> dgilmore: no I don't think so
14:47:12 <dgilmore> dustymabe: so you need to be explicitly clear
14:47:16 <dustymabe> dgilmore: basically what I want is all the config at the top that is common for them all to be in a global config file
14:47:24 <dgilmore> because I do not understand what change you are trying to make
14:47:30 <dgilmore> other than use includes
14:47:37 <dgilmore> I am not sure what you mean by that
14:47:57 <dgilmore> dustymabe: only osme of it is common
14:48:00 <dustymabe> ok, if you diff the config files in say fedora 25 branch
14:48:10 <dustymabe> you'll see mostly common configuration and then you'll see some tweaks
14:48:26 <dustymabe> i'm proposing that the common configuration be somewhere in a common config file
14:48:29 <dustymabe> that is included
14:48:34 <masta> dustymabe, I thinking that right now the atomic kickstart configs need to be separate
14:48:49 <dgilmore> mboddu: they are
14:49:05 <dgilmore> opps sorry mboddu
14:49:14 <dgilmore> masta: the kickstarts are seperate
14:49:19 <mboddu> dgilmore: :)
14:49:22 <dgilmore> words seem to be failing us
14:49:45 <dgilmore> because I read what masta said as the kickstarts have to be seperate
14:49:53 <dgilmore> which they already are
14:50:15 <masta> yes
14:50:21 <dustymabe> ok from my last set of statements, do you get what I'm saying now?
14:50:46 * masta goes to read the kickstarts for Fedora Atomic
14:50:59 <dustymabe> masta: this really has nothing to do with kickstarts
14:51:25 <masta> okay, my bad... pls help me understand?
14:51:37 <dustymabe> pungi configs are separate from kickstarts
14:51:54 <masta> yep
14:51:55 <dustymabe> a pungi config might define an image build and specify a kickstart to use
14:53:26 <dustymabe> ok - so from my previous set of statements
14:53:37 <dustymabe> i would propose that we take what we can from the top of the fedora.conf file
14:53:47 <dustymabe> and make it into a common config file for the others to #include
14:54:05 <dustymabe> here is an example of the parts of the file I would like to do that with: https://da.gd/FCDg
14:54:14 <dgilmore> dustymabe: kinda
14:54:24 <dgilmore> dustymabe: I think showing us what you mean might be easier
14:54:49 <dustymabe> so basically nothing that actually tells pungi to do an image-build or installer build or any build really
14:55:05 <dustymabe> just all the nitty gritty config
14:56:22 <dustymabe> dgilmore: see the fpaste for an example
14:56:56 <dgilmore> dustymabe: we should discuss it out of the meeting
14:57:07 <dustymabe> dgilmore: i tried that on the mailing list
14:57:12 <masta> Oh! this is about using %include in pungi, we do that all the time... so long as it validates via pungi-config-validate.. I'm +1
14:57:30 <dustymabe> masta: thanks - please add comments to the issue
14:57:41 * masta does
14:57:41 <dgilmore> masta: we do not do it at all
14:57:48 <dgilmore> masta: in Fedora
14:57:59 <dgilmore> anyway lets move on
14:58:05 <masta> dgilmore, oh! so I stand corrected. my bad
14:58:05 <dustymabe> dgilmore: well that's not true, there is at least one case where we already do
14:58:15 <contyk> let's talk about Xinclude in variants files ;)
14:58:23 <mboddu> dgilmore: lsedlar said that they are doing it internally
14:58:42 <mboddu> Okay, lets move on
14:58:45 <dustymabe> i don't know why we have to move on already - there's no resolution to this?
14:58:58 <dustymabe> are we saying "no" are we saying "maybe"
14:59:15 <dgilmore> dustymabe: we are not going to dolve it in the meeting and we have other things to discuss
14:59:23 <dgilmore> I think we should look at making things better
14:59:32 <dustymabe> well i tried to talk about this outside of the meeting
14:59:35 <dgilmore> but spending the whole meeting trying to do so is not productive
14:59:55 <mboddu> dustymabe: also, we need the documentation before we make any decision
14:59:58 <masta> dustymabe, the it has to validate pungi-config-validate or else nothing, I can say that with truth.. and that is trivial. so lets move on
15:00:32 <dustymabe> mboddu: thanks for bringing this up
15:00:36 <dustymabe> let's move on
15:00:49 <mboddu> #info masta will update the ticket and lsedlar will try to update the documentation on how to use %include
15:02:15 <mboddu> #topic #6791 Create module-bootstrap-master tag similar to f26-modularity
15:02:23 <mboddu> #link https://pagure.io/releng/issue/6791
15:02:56 <contyk> this one's been implemented, although the tag name is different than originally requested
15:02:59 <contyk> that's not a problem
15:03:32 <masta> not a problem
15:03:47 <dgilmore> contyk: rawhide seemed better to me that master
15:03:48 <contyk> we've also populated it with some content, with mboddu's help
15:03:50 <mboddu> contyk: okay, I just wanted to know if any more requests will come along on this issue? Like the secure-boot pkgs tagging and any others
15:04:10 <contyk> mboddu: yes, it might happen, over time
15:04:25 <contyk> mboddu: but if it has to be done, I'll open a new ticket
15:04:39 <contyk> as of now there's nothing else to be done
15:04:51 <dgilmore> I think this can be closed now
15:04:54 <contyk> +1
15:05:11 <mboddu> contyk: okay, just wanted to check, thanks
15:05:49 <mboddu> #info Nothing else can be done on this ticket as of now, if anything else pops up modularity-wg will open a new ticket. The ticket can be closed.
15:06:31 <mboddu> #topic #6775 Please Review Arbitrary Branching Change Proposal and Provide Feedback
15:06:38 <mboddu> #link https://pagure.io/releng/issue/6775
15:06:54 <mboddu> dgilmore: would you like to share any updates from FESCo?
15:09:00 * mboddu checks whats the new timeline for this
15:09:08 <dgilmore> FESCo acked it
15:09:22 <dgilmore> and said that the removal of pkgdb needs to be handled carefully
15:09:35 <dgilmore> it has been delayed a couple of weeks
15:10:17 <mboddu> dgilmore: so, is it still removing pkgdb and replacing it with new system, both being done at the same time?
15:11:48 <maxamillion> mboddu: "both being done at the same time" ... what is "both" here?
15:12:06 <mboddu> maxamillion: removing pkgdb and implementing new system
15:12:41 <mboddu> maxamillion: last time when we talked about it, I suggested using both like pkgdb for rpms and new system for modules and stuff
15:13:16 <maxamillion> mboddu: ohhh ok
15:13:24 <dgilmore> mboddu: the claim is that it can not be phased in
15:13:26 <mboddu> maxamillion: my suggestion was to use both of them for sometime and slowly move to new system when we feel like its good enough
15:13:34 <dgilmore> I suggested phasing in also
15:13:46 <maxamillion> I worry about consistency across systems
15:13:56 <masta> lets assume that removing pkgdb will not happen this cycle. hope for best, assume the worst.
15:14:07 <dgilmore> I need to follow up to some of the thread on devel@ about the packager group
15:14:32 <dgilmore> masta: its scheduled for remoaval in about 4-5 weeks
15:14:54 <mboddu> dgilmore: thats what scares me a lot
15:15:26 <dgilmore> mboddu: you and me both
15:15:52 <dgilmore> but all we can do is be engaged and make sure its a featureful as possible
15:16:00 <dgilmore> and raise concerns over issues we see
15:16:08 <mboddu> dgilmore: sure
15:16:31 <dgilmore> then make sure when its done we report issues and brokenees to the people working on it who are on the hook to fix broken things
15:16:56 <mboddu> #info FESCo acked the change and said that the removal of pkgdb needs to be handled carefully
15:16:56 <masta> wow
15:17:12 <mboddu> thanks dgilmore
15:17:26 <masta> dgilmore, understood.
15:17:40 <mboddu> moving on
15:17:43 <mboddu> #topic #6365 "Broken dependencies" report does not understand rich dependencies
15:17:50 <mboddu> #link https://pagure.io/releng/issue/6365
15:18:27 <mboddu> dgilmore: correct me if I am wrong, I thought we are planning on supporting dnf for f27 but not rich dependencies, am I right?
15:19:04 <mboddu> dgilmore: or are we going to support both?
15:21:14 <masta> uh...
15:21:39 <masta> so creterepo_s in mash, or koji dist-repos?
15:21:53 <masta> err... createrepo_c, even
15:22:12 <mboddu> masta: the plan was to port mash since koji dist-repos has some issues
15:22:37 <maxamillion> masta: I was going to ask what createrepo_s was, but was trying to google around first :)
15:22:54 <maxamillion> masta: I switched back to irc and noticed it was a fools errand
15:22:57 <contyk> S is for speed
15:23:04 <masta> it's a trivial fix, and I've had a patch in paguer for ages...
15:23:18 <maxamillion> masta: oh? is there a PR?
15:23:28 <masta> yes
15:23:39 <dgilmore> masta: correct
15:24:25 <dgilmore> masta: using the API is prefered
15:24:31 <dustymabe> i'm guessing the issues with dist-repos are not easily overcome
15:24:50 <dgilmore> dustymabe: dist-repos as is, is not functional
15:25:14 <dgilmore> main issue is that the debuginfo rpms are mixed in with the regular rpms
15:26:24 <dustymabe> k
15:27:02 <mboddu> Okay, lets move on
15:27:15 <mboddu> #topic Open Floor
15:27:58 <mboddu> #info Fedora 26 Beta got delayed by a week, current proposed release date is June 13 2017
15:28:13 <masta> maxamillion, https://pagure.io/mash/pull-request/1
15:28:14 <mboddu> Anybody got anything else?
15:28:16 <maxamillion> masta: +1
15:28:25 <maxamillion> no I
15:28:47 * mboddu got another meeting in 2 min
15:29:18 <mboddu> Okay, lets end the meeting
15:29:20 <dustymabe> mboddu:
15:29:31 <dustymabe> mboddu: ooops
15:29:32 <contyk> already? :(
15:29:32 <mboddu> dustymabe: you got anything?
15:29:33 <dustymabe> sorry
15:29:35 <dustymabe> nope
15:29:38 <mboddu> dustymabe: okay
15:29:45 <contyk> could I get one more topic in?
15:30:04 <contyk> https://pagure.io/releng/issue/6815
15:30:32 <contyk> the only releng task associated with this change has been implemented
15:30:45 <contyk> I was wondering if I could get a releng ack on the change so I can push it forward
15:30:49 <mboddu> contyk: we are over our time and I got another meeting now, can we take this to the #fedora-releng channel
15:31:04 <mboddu> contyk: sorry about it
15:31:26 <dgilmore> contyk: been meaning to do that, I need to coordinate with katec to make sure we get any work needed in our backlog
15:31:44 <contyk> dgilmore: thanks
15:31:55 <mboddu> so, we got everything
15:32:01 <mboddu> Thanks everyone for joining
15:32:05 <mboddu> #endmeeting