15:00:23 <pknirsch> #startmeeting Fedora Base Design Working Group (2014-02-14)
15:00:23 <zodbot> Meeting started Fri Feb 14 15:00:23 2014 UTC.  The chair is pknirsch. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:36 <pknirsch> #meetingname  Fedora Base Design Working Group
15:00:36 <zodbot> The meeting name has been set to 'fedora_base_design_working_group'
15:01:07 <pknirsch> good morning and good afternoon everyone (or in whatever timezone you might be right now)!
15:01:49 <pknirsch> i already saw masta :)
15:01:52 <pknirsch> #chair masta
15:01:52 <zodbot> Current chairs: masta pknirsch
15:01:54 * masta is here
15:02:00 * jreznik is here but a bit semi present
15:02:06 <pknirsch> !
15:02:10 <pknirsch> #chair jreznik
15:02:10 <zodbot> Current chairs: jreznik masta pknirsch
15:02:16 <pknirsch> aren't we all, jreznik :)
15:02:20 <jreznik> pknirsch, are you feeling better?
15:02:26 * pknirsch is still recovering from the DevConf plague
15:02:37 <pknirsch> yea, thankfully i managed to sleep most of it off yesterday
15:02:51 * jreznik got hiy with that con plague too
15:02:54 * dgilmore is here
15:03:00 <pknirsch> hey dgilmore :)
15:03:05 <pknirsch> #chair dgilmore
15:03:05 <zodbot> Current chairs: dgilmore jreznik masta pknirsch
15:04:04 <pknirsch> ok, lets get started, especially for the european folks to not prolong it too long into the Friday evening.
15:04:14 <dgilmore> pknirsch: im still recovering, being here still may not help
15:04:39 <pknirsch> dgilmore: yea, though the beer and food might help :)
15:04:50 <dgilmore> pknirsch: some days
15:05:06 <pknirsch> :)
15:05:34 <pknirsch> but seems a lot of people got hit with it, radek was out as well
15:05:59 <dgilmore> lots of it this week
15:06:13 <pknirsch> mhm
15:06:19 <pknirsch> ok, so first topic for today:
15:06:23 <pknirsch> #topic Cleanup status/report
15:06:33 <pknirsch> just wanted to give a quick update on that.
15:06:55 <pknirsch> First up, we got our first change due to my review: ca-certificats doesn't require java anymore for building, yay! :)
15:07:07 <dgilmore> :)
15:07:43 <pknirsch> due to the conf any me being out cold and tied to lots of other s**t this week i couldn't make much more progress otherwise, but i've looked at the remaining java buildrequires. the automake one is already gone, too in rawhide
15:08:02 <pknirsch> and the only one left is the libdb and libdb4 for the java bindings
15:08:32 <pknirsch> after running repoquery on f20 though there were only 2 packages in Fedora the require them: libdb-devel and libdb4-devel :)
15:08:38 <masta> we should create badge for packagers, removed BR or R from package, reduce the dep chains
15:08:46 <masta> *joking*
15:08:46 <pknirsch> good idea!
15:08:49 <pknirsch> :P
15:09:22 <pknirsch> i'll send out an email to f-d-list to see if there is anyone using them for their own stuff
15:09:37 <pknirsch> and if not i'll ask the maintainer if we could disable the build for the bindings for db4
15:09:48 <masta> pknirsch: I'm glad we are starting to see things reduce
15:10:06 <pknirsch> that should get us java out of the self-hosting build cloud
15:10:11 <pknirsch> masta: me too!
15:10:34 <dgilmore> pknirsch: BuildRequires: libdb-devel
15:10:37 <dgilmore> from rpm
15:10:56 <masta> ah, so db4 pulled in java... wow
15:11:06 <masta> yea... that is wacky
15:11:19 <pknirsch> but does it require the java part dgilmore ? but good point, didn't check the transitive deps yet
15:12:01 <dgilmore> pknirsch: ahh libdb-java-devel no
15:12:12 <pknirsch> ah, good :)
15:12:18 * dgilmore was having a blond moment
15:12:24 <pknirsch> well
15:12:27 <pknirsch> :)
15:12:50 <pknirsch> we all have, but still good to have it double checked, thanks!
15:14:00 <pknirsch> as mentioned before i've done some tests with rwmjones autobuildrequires and it worked quite nicely, but haven't really investigated deeply into how to use the info from that and crosscheck it with the specfile and generate a report of the diff
15:14:11 <pknirsch> but that should be a days' worth of hacking at most
15:15:39 <pknirsch> i've talked on devconf with several folks about pregenerating the docs. most were curious as to how to do it and voiced some concerns about more effort or side effects, so i'll see if i can make a few examples of how to do it and document it on the wiki
15:15:47 <pknirsch> but that'll be a few weeks out at least
15:16:21 <masta> that is the script that iterates overs BRs, removes BR then builds, and notes when it reaches a build failure?
15:16:57 <masta> right?
15:17:01 <jreznik> for docs, it's a good idea to show the way on wiki
15:17:14 <pknirsch> that'll be the final stage of the autobuildrequires combo, masta
15:17:18 <pknirsch> at least thats the plan
15:17:32 <pknirsch> but that might take a few more days of hacking as thats not as trivial as i initially assumed
15:17:48 * pknirsch didn't mention "how hard can it be..."
15:18:03 <pknirsch> jreznik: aye. and provide some typical scenarios there if possible
15:18:26 <pknirsch> e.g. how to pre generate docs if your software used doxygen for generating them etc
15:18:28 <masta> pknirsch: sounds like a fun script
15:18:33 * pknirsch nods
15:19:06 <pknirsch> if you want to twiddle with the idea masta, by all means do so :) always looking for volunteers ;)
15:21:14 <masta> pknirsch: I've got something here that I used for doing mock builds, that could probably be adapted... I'll independently give it a shot as time allows =)
15:21:28 <pknirsch> cool, thanks!
15:22:21 <pknirsch> #info First dep removal: ca-certificates no longer requires java to build
15:22:52 <pknirsch> #info libdb and libdb4 only ones left requiring java for building
15:23:31 <pknirsch> #action pknirsch send out email to f-d ML to check if anyone is using the bindings. If not, contact maintainer to see if those could be disabled for Fedora
15:23:47 <pknirsch> #info Rest of the items still work in progress
15:24:05 <pknirsch> alright, lets move over to DevConf quickly
15:24:10 <pknirsch> #topic DevConf meet up summary
15:24:34 <pknirsch> so dgilmore, jreznik, me and tons of other folks were at DevConf in Brno last Friday-Sunday
15:24:49 <pknirsch> Numerous talks about docker docker docker and, docker. :)
15:24:52 <jreznik> and got sick there ;-)
15:24:57 <pknirsch> true :)
15:25:16 <pknirsch> they should rename Inflentia to Confluentia
15:25:55 <dgilmore> :D
15:26:13 <pknirsch> Had a blast there, tons of chats with folks and the WG representatives did a 1 hour Q&A session there.
15:26:16 <jreznik> For docker and company - do we want/have to do something in base?
15:26:29 <jreznik> also os tree was mentioned there pretty often
15:26:39 <dgilmore> is it something all working groups will want?
15:26:53 <dgilmore> I know cloud wants it
15:26:56 <dgilmore> does any other?
15:27:30 <jreznik> cloud for sure, would be nice to check other groups and can serve very well for qa
15:27:38 <pknirsch> thats one of the things i wanted to discuss here, yes. i'm sure cloud wants it, and i can see Workstation potentially as well. But unsure about Server, but i could see a potential there for very odd roles
15:28:11 <pknirsch> But Workstation might just go with the systemd nspawn containers without the docker stuff around it
15:29:14 <pknirsch> I can ping the other WG reps by next week to see what if they want/need docker for their product
15:29:16 <pknirsch> same for ostree
15:30:00 <dgilmore> well ostree is not suitable as a distribution method
15:30:12 <dgilmore> it has some very good and valid use cases
15:30:28 <dgilmore> but for distributing a distribution not so much
15:31:02 <dgilmore> its much more suited to intra organisation distribution than Fedora distribution
15:31:13 <pknirsch> how so? i thought thats one of the nice things about it, that you can simply check out a specific version of the tree you'd like to run and then boot into it, right?
15:31:22 <dgilmore> but we are going to look at doing some ostree stuff for qa use
15:31:27 <pknirsch> ah
15:31:45 <dgilmore> well its thousands of files. mirrors will hate it
15:31:58 * pknirsch nods
15:32:03 <dgilmore> you would need lots and lots of different trees
15:32:14 <masta> yea... we want to keep the mirrors happy
15:32:18 <jreznik> and if it's going to be available for qa and cloud, then it's base and we should own it
15:32:19 <dgilmore> and from a user perspective you get no choice in hsats in the tree
15:33:05 <dgilmore> but in say a company workstation deployment, app servers, proxy servers, etc then its definetly a good option
15:33:19 <pknirsch> right, no updates or modifications in the tree was one problematic point. the other one was the need for an update when you want to switch to another version of a tree
15:33:28 * pknirsch nods
15:33:33 <pknirsch> jreznik: good point
15:34:04 <dgilmore> jreznik: ostree trees are not going to be available for cloud
15:34:23 <pknirsch> didn't cloud want to distribute images?
15:34:35 <dgilmore> images are not ostree trees
15:34:39 <pknirsch> right
15:34:44 <dgilmore> they want docker images
15:34:52 <dgilmore> and appliance images
15:35:08 <pknirsch> thats why i'm asking, yea. so ostree would only be needed for CI work of QE or other use cases right now?
15:35:19 <masta> If I am to imply that OStree's are check-points of our fedora... a type of acyclic-graph... then that would be great for QA work
15:35:28 <pknirsch> aye
15:35:39 <dgilmore> things that are needed to compose fedora all fall into Base by the definition we got on week 1
15:35:39 <pknirsch> for finding bugs this is huge!
15:35:44 <pknirsch> aye
15:36:22 <pknirsch> but distributing is a different matter, totally agree dgilmore
15:36:35 <dgilmore> which will make the tooling for docker images, ostree (even for internal only use), any other type of thing that comes up all fall in base
15:36:46 <pknirsch> yea
15:37:02 <jreznik> well, I understood mattdm would like to use OSTree for cloud, but maybe I was confused by something different
15:37:11 <pknirsch> hm
15:37:19 <mattdm> jreznik I'm interested in exploring it. :)
15:37:19 <dgilmore> probably should make a releng-tools group in comps
15:37:35 <dgilmore> or someother way to manage whats pulled in for compose reasons
15:38:55 <pknirsch> you can explore colin's images today, mattdm :)
15:39:28 * dgilmore really needs to start a conversation with mirrors
15:39:38 <dgilmore> and see what types of things they will accept
15:40:11 * pknirsch nods
15:40:12 <masta> dgilmore: it would be nice to have some expansion
15:41:39 <pknirsch> and i'd think especially QE could give us a good indicator when they use ostree how stable and usable it is. As if we'd distribute Fedora that way and people would have a bad experience with that that would suck
15:42:01 <dwalsh> I see docker as being important to Server and cloud.  As of right now not that important to Workstation.
15:42:24 <pknirsch> ah, for the gazillions of httpd server for server, right, i remember dwalsh
15:42:38 <dwalsh> Workstation is looking at using a different kind of containers.  Using part of the Base OS as well as bundeled app.
15:42:45 * pknirsch nods
15:43:01 <pknirsch> kinda sad they can't use docker. otherwise we'd use the same tech in all 3 products
15:43:10 * dwalsh attending two meetings simultaneously...
15:43:16 <pknirsch> :)
15:43:46 <dwalsh> pknirsch, I think they will end up using docker at desktop also, since third parties might just package desktop app in docker.
15:44:02 <pknirsch> right
15:44:06 <dwalsh> But it is very early in the Desktop Apps world.
15:44:10 <pknirsch> thats what i'm hoping
15:45:07 <pknirsch> and just to be clear, the tooling would reside in Base, but the actual images would be product specific i think, right dgilmore ?
15:46:14 <dgilmore> pknirsch: right
15:46:19 <jreznik> and docker is going be good friend of systemd
15:47:12 <masta> well the nspawn stuff is pretty much free
15:47:19 <pknirsch> exactly
15:47:32 <masta> so the notion of workstation being different is kinda weird
15:47:45 <masta> moe like it just opts out of an enhancement
15:47:49 <jreznik> yep
15:47:58 <masta> but if the other two charge forward that is an incentive to them
15:48:13 <pknirsch> #info docker and ostree tooling to be available in Base to due rel-eng dependency
15:48:49 <jreznik> owned
15:49:52 <masta> yea... I'd say it is more imperative.. owned, or required, mandantory, must...etc
15:49:52 <pknirsch> right
15:51:30 <pknirsch> ok, but lets quickly jump to the last topic, running out of time already again :/
15:51:36 <pknirsch> #topic Requirements/changes Base needs in Fedora (https://fedorahosted.org/fesco/ticket/1178)
15:52:08 <pknirsch> From the FESCO meeting on Wednesday and the ticket though fesco wants to provide more guidance to the WGs whats expected from them
15:52:50 <pknirsch> not sure if we have a clear set of requirements yet for Base though. thats probably mainly a question to dgilmore and masta as it'll mostly revolve around rel-eng i suspect
15:53:21 <dgilmore> pknirsch: yeah, I can try write up what I think we will need
15:54:07 <pknirsch> excellent, thanks dgilmore.
15:54:09 <jreznik> And also we are in the chain of other groups
15:54:50 <pknirsch> i've wrecked my clogged brain a bit the last days, but i didn't see any other requirements from Base coming out to other groups or WGs
15:55:20 <jreznik> Not now and I think it should be part of that initial scoping
15:55:45 <pknirsch> exactly jreznik
15:55:47 <jreznik> we should be more proactive here, let me try to express it to other WGs
15:56:11 <pknirsch> i see other WGs requirements trickle down and either land in Base, rel-eng, QE, docs or somewhere else
15:56:27 <jreznik> yep
15:57:05 <jreznik> But they should be aware we are waiting for input - mattdm, Jwb & sgallagher ;-)
15:57:22 <pknirsch> :)
15:57:36 <pknirsch> i've already taken a look at the tech spec of the workstation WG
15:57:57 * mattdm nods, takes a note
15:58:05 <pknirsch> it's pretty detailed and thorough, i'll try to see if i can extract some more info for Base till next week
15:59:18 <jreznik> pknirsch, well it's really draft, not voted on and I expect pretty heated discussion...
15:59:37 <pknirsch> sure
15:59:46 <pknirsch> but it's a good starting point to extract some info
16:00:11 <pknirsch> as the choice of KDE or not won't really affect Base much i think ;)
16:00:18 * dgilmore feels he needs to spend the next week reading
16:00:39 <dgilmore> pknirsch: porting anaconda to qt5 would
16:00:53 <dgilmore> but i dont think thats being proposed or considered
16:01:19 <pknirsch> it isn't?
16:01:25 * pknirsch is shocked!
16:01:33 <jreznik> one of anaconda developers is silently thinking about it ;-)
16:01:35 <dgilmore> maybe I just planted the seed
16:01:41 <pknirsch> hahahah
16:01:47 * nirik has a random question that might be answered already...
16:01:55 <dgilmore> nirik: sure
16:01:55 <pknirsch> shoot nirik !
16:01:58 <pknirsch> pew
16:02:22 <nirik> is base going to be producing/managing the boot.iso? or having any deliverable? or is boot.iso going to be more a server thing?
16:02:36 <pjones> dgilmore: ha ha funny no.
16:02:43 * nirik thinks base making a deliverable would help be able to test it/etc.
16:03:12 <pknirsch> i think we agreed to provide the netboot at least iirc
16:03:17 <pknirsch> that was a while ago though
16:03:25 <dgilmore> nirik: thats on the cards
16:03:39 <dgilmore> nirik: boot.iso and netinstall tree for installer testing
16:03:42 * pknirsch nods
16:03:43 <nirik> ok, will that be the current one? or your 'self hosting set' ? :) (not sure how big that would be)
16:04:05 * jreznik has to quit now
16:04:07 <nirik> it would be kinda cool to have a self hosting thing on one iso...
16:04:18 <dgilmore> nirik: not yet 100% sure
16:04:34 <nirik> ok. Thought I would ask
16:04:59 <dgilmore> likely it will be a Base repo and set of installers
16:05:30 <dgilmore> where Base repo is the self hosting base or installer, minimal environment and compose tools
16:05:39 <dgilmore> s/or/of/
16:05:41 * pknirsch would love that, too
16:07:07 <pknirsch> ok, lets put down an action for dgilmore for the writeup
16:08:11 <pknirsch> #action dgilmore to put together a requirement document for rel-eng/Base
16:08:36 <pknirsch> any other important topics for today?
16:09:31 <pknirsch> if not, thanks everyone for joining today and see you next week, hope we can tackle the other 2 topics that were on the optional list for today then,
16:09:42 <pknirsch> have a great weekend everyone! and hope evey
16:09:44 * masta waves bye
16:09:54 <dgilmore> pknirsch: enjoy the weekend
16:09:57 <pknirsch> ryone who was at DevConf and got the plague can recover a bit over the weekend
16:10:10 <pknirsch> o/
16:10:15 <pknirsch> #endmeeting