From Fedora Project Wiki
(09:59:55 AM) thl has changed the topic to: FESCo meeting -- init
(09:59:57 AM) ***mmcgrath is here
(10:00:02 AM) ***warren here
(10:00:07 AM) ***bpepple is here.
(10:00:08 AM) ***c4chris here
(10:00:13 AM) ***awjb is here
(10:00:20 AM) thl: hi everyone
(10:00:24 AM) tibbs: I'm here.
(10:00:27 AM) dgilmore: heya ll
(10:00:35 AM) thl: warren, I had planed that in any case
(10:00:43 AM) ***nirik is in the rabble seats
(10:00:46 AM) warren: k, be back in a few minutes
(10:00:53 AM) thl: that let's start with the easy stuff
(10:01:00 AM) thl has changed the topic to: FESCo meeting -- EPEL
(10:01:14 AM) thl: the discussion was quite confusing last week
(10:01:18 AM) thl: so where are we now
(10:01:23 AM) thl: what is out next step?
(10:01:26 AM) thl: mmcgrath, dgilmore ?
(10:01:27 AM) ***cweyl is here (rabble)
(10:01:29 AM) bpepple: Are there any minutes from last week?
(10:01:45 AM) ***thl tries to remember
(10:01:45 AM) dgilmore: thl: we have had some builds
(10:01:53 AM) thl: bpepple, no, I don't think so
(10:01:58 AM) bpepple: ok.
(10:02:02 AM) tibbs: bpepple: I have logs, but I didn't see any minutes.
(10:02:04 AM) dgilmore: thl: just test stuff
(10:02:13 AM) mmcgrath: dgilmore: what did we do with the ppc kernel issue?
(10:02:16 AM) dgilmore: i need to do some more branching and test some other builds
(10:02:23 AM) ***abadger1999 has arrived
(10:02:31 AM) dgilmore: mmcgrath: im going to put a kernel in there right now
(10:02:41 AM) mmcgrath: k.
(10:02:45 AM) thl: dgilmore, mmcgrath, I'd like it if one of you two (or both?) could be responsible for EPEL and driing it furhter
(10:02:53 AM) thl: further
(10:03:02 AM) mmcgrath: thl: AFAIK we're about ready start building, getting policies together and actually publishing.
(10:03:09 AM) dgilmore: thl: sure we can work together on it
(10:03:38 AM) dgilmore: thl: one  thing we really need is to know where we are going to put the final signed packages
(10:03:49 AM) thl: dgilmore, that mean driving the whole thing foward (just to make sure you know what I'm talking about)
(10:04:00 AM) thl: dgilmore, there are some ideas in the wiki
(10:04:14 AM) dgilmore: thl  i understand
(10:04:21 AM) ***rdieter is back from never never land.
(10:04:25 AM) thl: dgilmore, if you own it then you should find the solutions
(10:04:30 AM) thl: and just present them to us
(10:04:48 AM) thl: and we'll probably say "yes, that nice" most of the time
(10:04:50 AM) dgilmore: thl: the main issue is do we put in it /pub/fedora somewhere or /pub/epel/
(10:05:01 AM) dgilmore: or some other seperate from fedora tree
(10:05:11 AM) dgilmore: thl: ok
(10:05:13 AM) thl: my ote is
(10:05:30 AM) thl: (my "v" still makes problems)
(10:05:31 AM) cweyl: it's a fedora project under the jurisdiction of fesco, right?  so wouldn't it be...  what thl just said :)
(10:05:46 AM) mmcgrath: we can continue this on the list.
(10:05:58 AM) thl: mmcgrath, thx, that would be great
(10:06:03 AM) ***mmcgrath suspects there's lots of summit stuff to talk about ;-)
(10:06:07 AM) rdieter: the only argument for a separate tree is that it lets mirrors opt out more easily.
(10:06:07 AM) thl: mmcgrath, dgilmore, and please keep the wiki up2date
(10:06:31 AM) thl: well, and might go away sooner or later, too
(10:06:42 AM) dgilmore: rdieter: the argument i heard for a summit tree is some enterprise people will be confused if its in fedora space
(10:06:44 AM) thl: so putting epel below there would probably not be that wise
(10:06:50 AM) thl: anyway
No such command.
(10:06:53 AM) thl: let's stop here
(10:06:55 AM) thl: and move on
(10:06:57 AM) rdieter: dgilmore: tough, I say. (:
(10:07:07 AM) scop [n=scop]  entered the room.
(10:07:08 AM) thl: we probably still hae lot's to discuss
(10:07:26 AM) thl has changed the topic to: FESCO meeting -- MISC, topic 1
(10:07:33 AM) thl has changed the topic to: FESCO meeting -- MISC, priority 1
(10:07:47 AM) thl: did anybody read ?
(10:08:00 AM) ***rdieter nods, looked good.
(10:08:21 AM) ***thl waits for further opinions
(10:08:32 AM) tibbs: Yes, this looks good.
(10:08:33 AM) ***cweyl skims quickly
(10:08:34 AM) abadger1999: I scanned it quickly.  It looked fine
(10:08:39 AM) bpepple: looks good to me.
(10:08:58 AM) c4chris: fine with me too
(10:09:01 AM) awjb: if it did not change since lunch it looks good :)
(10:09:11 AM) thl: k, then let's make them officical?
(10:09:15 AM) thl: some +1 please
(10:09:15 AM) dgilmore: looks ok to me
(10:09:17 AM) dgilmore: +1
(10:09:20 AM) cweyl: thl: I like it.  it doesn't mention explicitly the traditional "vote by screaming", but I like :)
(10:09:38 AM) abadger1999: +1
(10:09:40 AM) tibbs: Did this change since you first posted it?
(10:09:43 AM) c4chris: +1
(10:09:49 AM) thl: tibbs, nope, I don#t think so
(10:09:51 AM) awjb: +1
(10:09:52 AM) bpepple: +1
(10:09:54 AM) tibbs: Looks like it didn't.  So +1.
(10:09:55 AM) rdieter: +1
(10:10:09 AM) thl: k, settled then
(10:10:23 AM) thl: we can still adjust it later if we want / hae to
(10:10:25 AM) thl: have
(10:10:40 AM) thl has changed the topic to: FESCo meeting -- Packaging Committee Report
(10:10:47 AM) thl: was there anything to report?
(10:11:01 AM) tibbs: No, many members off to the summit.
(10:11:04 AM) abadger1999: Not enough people for any decisions
(10:11:12 AM) thl: k, then lets move on
(10:11:17 AM) thl has changed the topic to: FESCo meeting --  Sponsorship nominations
(10:11:35 AM) thl: there were no further discussion after last weeks meeting
(10:11:46 AM) thl: do we want to discuss the current nominations again?
(10:11:54 AM) thl: Steven Pritchard and Paul F. Johnson ?
(10:12:03 AM) thl: or leave that for next week?
(10:12:22 AM) tibbs: We really need to address the criteria.
(10:12:23 AM) abadger1999: I think we voted yes on silug last week
(10:12:32 AM) abadger1999: And were going to discuss pfj further.
(10:12:47 AM) tibbs: I don't have any objections to either of them as people, and think they'd both do a fine job.
(10:13:02 AM) thl: okay, I'll upgrade silug to sponsor status if nobody yells in the next minute
(10:13:09 AM) tibbs: But Steven doesn't do reviews.  If that's a problem, then we need to discuss it.
(10:13:27 AM) thl: tibbs, I don#t think it's a big problem
(10:13:43 AM) thl: tibbs, I also did not do to much reiews
(10:13:47 AM) rdieter: no reviews, not a blocker (nice, but not essential, imo).
(10:13:49 AM) thl: abd became sponsor, too
(10:13:59 AM) tibbs: Who is abd?
(10:14:16 AM) thl: abd?
(10:14:25 AM) scop: s/abd/and/?
(10:14:36 AM) thl: sorry, typo
(10:14:38 AM) tibbs: Ah, sorry.
(10:14:43 AM) warren: back, that took longer than expected. =(
(10:14:51 AM) tibbs: OK, +1 for Steve.
(10:15:06 AM) thl: k, I'll upgrade him later
(10:15:16 AM) thl: just for completeness
(10:15:20 AM) tibbs: Can anyone who has an objection to Paul Johnson please state it somewhere so at least he knows what the issue is?
(10:15:30 AM) thl: bpepple, what was your opinion on making Paul F. Johnson ?
(10:15:36 AM) rdieter: no objections here.
(10:15:42 AM) thl: bpepple, I got a bit confused in hte last meeting (my fault)
(10:16:01 AM) bpepple: I'm sorta hesitant on him based on his past reviews.
(10:17:13 AM) thl: bpepple, I know what you mean
(10:17:25 AM) thl: I'm a bit unsure myself
(10:17:31 AM) thl: buit his sponsor gae his okay
(10:17:43 AM) thl: gave
(10:17:52 AM) thl: so he get's a +1 from me
(10:18:08 AM) thl: bpepple, is that more a "0" or more a "-1" ?
(10:18:21 AM) bpepple: More a '0' than a '-1'.
(10:18:23 AM) rdieter: +1 (learning from mistakes is a good thing)
(10:18:29 AM) thl: rdieter, agreed
(10:18:33 AM) thl: any more +1 ?
(10:18:33 AM) dgilmore: +1
(10:18:34 AM) tibbs: I don't expect anyone to be perfect, and from what I've seen he has continued to improve.
(10:18:37 AM) tibbs: +1
(10:18:38 AM) warren: +1
(10:18:42 AM) awjb: +1
(10:18:43 AM) abadger1999: I think pjf doesn't always understand the big picture but since the last time we reviewed him he's come to understand the details pretty well.
(10:18:48 AM) c4chris: +1
(10:19:01 AM) thl: abadger1999, agreed
(10:19:26 AM) thl: okay, I consider him accepted if nobody yells soon
(10:19:53 AM) thl has changed the topic to: FESCo meeting --  Merge Core and Extras
(10:19:57 AM) thl: warren, rdieter ?
(10:20:07 AM) warren: let's focus on specific parts of this one at a time
(10:20:19 AM) warren: first, has everyone read up on the summit stuff?
(10:20:24 AM) awjb: y
(10:20:25 AM) warren: especially my blog entry
(10:20:37 AM) abadger1999: yes
(10:20:40 AM) bpepple: I read a bit of it yesterday on the plane trip home.
(10:20:41 AM) scop: n
(10:20:41 AM) warren:
(10:20:44 AM) c4chris: in part
(10:20:51 AM) thl: warren, well, ther was much stuf to follow, but yes
(10:21:02 AM) warren: "Essentially all packages will become maintained in a manner similar to today's Extras project, but with more sophisticated controls, process and policy automation. Core will merge into Extras, creating one big distribution maintained by both RH engineering and Fedora contributors. The name for the new distribution is currently undecided, but may just be called "Fedora".
(10:21:02 AM) tibbs: I've been following the IRC channel.
(10:21:37 AM) warren: let's focus on the first topic... the review process
(10:21:45 AM) warren: "ll Core packages will go through a Packaging Guideline review before merging. An abbreviated review process was discussed briefly, and details still need to be worked out. A trickle of low-hanging fruit (usually leaf-node apps like squirrelmail or gaim that wont break deps) can begin with standard reviews sooner. Mass movement of packages to Extras is pending FESCo approval, where discussion begins during Thursday's regular meeting. I am per
(10:21:45 AM) warren: sonally accountable to this part of the plan."
(10:22:13 AM) warren: Mass review and movement of packages wont happen until February-ish
(10:22:35 AM) thl: how will this "mass review" look like?
(10:22:43 AM) warren: Until then, core maintainers have the option of moving "easy" packages using the current standard way, with a review process.
(10:22:53 AM) warren: Mass review we need to figure out the details between now and February.
(10:23:42 AM) thl: well, I have no problem with it if core wants to moe some packages over
(10:23:48 AM) thl: we did this in the past
(10:23:59 AM) ***nirik thinks thats going to be a balancing act... too easy and bad quality packages, too hard and they won't get done in time for FC7 (or whatever it's called)
(10:24:01 AM) thl: so why stop now
(10:24:02 AM) warren: Yeah, using standard review process, no problem.
(10:24:20 AM) c4chris: for one thing, the "check the license" point is done on all core packages
(10:24:20 AM) warren: nirik, maybe
(10:24:38 AM) nirik: also build in min buildroot should be done. ;)
(10:24:42 AM) jima: c4chris: good point, and that's always a PITA :)
(10:24:46 AM) c4chris: maybe there are already several points taken care of (like builds in mock)
(10:24:48 AM) thl: nirik, we might need to sort out some issues oer time (e.g. after FC7)
(10:25:05 AM) warren: There was one proposed mass review process during the summit, it might be too simplistic.
(10:25:08 AM) thl: nirik, core builds with the same buildroot as extras these days iirc
(10:25:13 AM) warren: We need to design a better and comprehensive process.
(10:25:21 AM) thl: nirik, so it should work out of the box afaik
(10:25:31 AM) nirik: it's much harder to go back and fix something once it's already in and working than if it needs to be fixed first. IMHO. :)
(10:25:41 AM) jima: thl: ask dgilmore, he's built much/most of core in plague.
(10:25:50 AM) warren: The discussed process was to just import everything into Extras, with a "" file.  Reviewers just remove that file with a checkin saying why they approve it.
(10:25:52 AM) jima: (for sparc, but...)
(10:25:57 AM) warren: The idea here was to avoid filing bugs for each one.
(10:26:00 AM) warren: I don't exactly like this plan.
(10:26:18 AM) thl: warren, does brew use the same minimal muildroots that extras uses?
(10:26:18 AM) abadger1999: arren: I'm with you on that one.
(10:26:24 AM) c4chris: why not open bz tickets ?
(10:26:28 AM) cweyl: minor pedantic note -- and I might be getting ahead of myself -- but for any of the 'easy migrations', would it be possible to track them with a blocker bug?
(10:26:28 AM) abadger1999: s/^/w/
(10:26:33 AM) rdieter: c4chris: 1000's of them? (:
(10:26:36 AM) bpepple: warren: agreed.
(10:26:37 AM) dgilmore: thl: they are supposed to be using the same minimal buildroot
(10:26:41 AM) tibbs: Another problem is that we have to consider upgrade comatibility with old releases, so we might not be as free to move things around so that everything is stored in the proper place.
(10:26:47 AM) warren: thl, good question, better check with f13 on that one.
(10:26:59 AM) c4chris: better and uniform tracking this way
(10:27:27 AM) warren: tibbs, Core moving to Extras (the trickle) between now and the mass review is only devel, and not meant to build a distro.
(10:27:37 AM) warren: maybe trickle is a bad idea?  I dunno.
(10:27:38 AM) thl: well, I don't think we hae to work out a "mass review" scheme today
(10:27:47 AM) thl: I think we should discuss this on the list
(10:27:51 AM) ***rdieter nods
(10:27:51 AM) bpepple: thl: +1
(10:27:56 AM) dgilmore: i like the trickle  idea
(10:27:56 AM) warren: thl, +1
(10:28:03 AM) c4chris: thl, +1
(10:28:12 AM) warren: OK, next major topic...
(10:28:17 AM) warren: What happens to FESCo?
(10:28:17 AM) thl: warren, can you bring it up for discussion there?
(10:28:18 AM) rdieter: trickle, raging torrent, whatever +1, the more the merrier I say
(10:28:39 AM) warren: We think nothing happens to FESCo.  Just the mandate grows larger.
(10:28:44 AM) jima: warren: they get assassinated by the fedora core cabal
(10:28:45 AM) thl: warren, well, what did the summit discuss?
(10:28:48 AM) warren: And more RH people have to pay attention to FESCo.
(10:28:48 AM) ***cweyl would still like to see them tracked with a blocker bug :)
(10:28:50 AM) bpepple: and possibly a name change.
(10:29:05 AM) thl: warren, well, what did the summit discuss (regaring the future of FESCO)?
(10:29:10 AM) warren: Yeah, we need a new name for FESCo.  That will probably rely on whatever name is chosen for Fedora distro itself.
(10:29:43 AM) warren: thl, FESCo remains much the same in what it does, except it becomes more powerful.
(10:29:44 AM) rdieter: Fedore Communityr Steering Comittee, FCSCO (you can sound out the anacronim for yourself) (:
(10:29:48 AM) warren: Muwhahahhaha
(10:29:52 AM) tibbs: But what is the committee's place?
(10:30:04 AM) c4chris: rdieter, :-)
(10:30:13 AM) tibbs: Does FAB run the show? What happens to the core cabal?
(10:30:15 AM) thl: yeah, But what is the committee's place?
(10:30:16 AM) rdieter: Same job, bigger scope.
(10:30:26 AM) abadger1999: tibbs: Technical decisions?
(10:30:26 AM) thl: rdieter, do we need more poeple for it?
(10:30:41 AM) abadger1999: So we are in effect merged with the Core Cabal?
(10:30:42 AM) warren: If all packages are put into Extras, and the project is run in a very Extras-like way...
(10:30:46 AM) rdieter: thl: maybe, personally, I don't think so...
(10:30:54 AM) warren: abadger1999, err... I don't know.
(10:31:01 AM) tibbs: abadger1999: Are you suggesting letting FAB handle politics while FESCO2 handles technical stuff?
(10:31:02 AM) warren: Yeah, we don't really know all implications just yet.
(10:31:07 AM) warren: Things will become more complicated.
(10:31:23 AM) warren: tibbs, indeed, FAB handles politics primarily
(10:31:25 AM) rdieter: tibbs: sounds reasonable.
(10:31:27 AM) dgilmore: i thoght the core cabal  was really just jeremy, f13 , and notting
(10:31:42 AM) rdieter: dgilmore: yup.
(10:31:55 AM) jima: dgilmore: that's what "they" want us to think. ;)
(10:32:06 AM) abadger1999: We already have jeremy on FESCo :-)
(10:32:06 AM) [splinux]  left the room (quit: "Ex-Chat").
(10:32:20 AM) dgilmore: abadger1999: yeah so it would be two additional people
(10:32:36 AM) warren: from a technical perspective, FESCo would be very similar, except by necessity we would have to work much closer to RH engineering.
(10:32:48 AM) jeremy: sorry, forgot that it was thursday
(10:32:50 AM) dgilmore: i say bring them in the have alot of knowledge of things that we dont
(10:33:02 AM) ***dgilmore slaps jeremy   wake up :D
(10:33:03 AM) warren: Thus I suspect we might need *ONE* representative from each major RH engineering team on FESCo.  For example, we currently don't have the desktop team paying attention to us.
(10:33:04 AM) ***thl hits jeremy with a calendar
(10:33:15 AM) rdieter: jeremy: tis ok, I forgot what day it was too...
(10:33:16 AM) jeremy: it feels like monday, sort of!
(10:33:38 AM) iWolf_ [n=jft]  entered the room.
(10:33:43 AM) warren: Jeremy already represents installer/OS very well.
(10:33:43 AM) dgilmore: jeremy, rdieter:  thats cause you guys goofed off the rest of the week :D
(10:33:51 AM) warren: Jesse represents release engineering
(10:33:52 AM) rdieter: warren: makes sense.  we should have more rh reps.
(10:34:00 AM) warren: I represent (random other things)
(10:34:01 AM) thl: warren, does the board or red hat want a fixed ratio "redhat vs community" members for FESCo in the longer term?
(10:34:14 AM) warren: thl, we hadn't discussed that specifically
(10:34:14 AM) rdieter: (not too many, we still need to be able to take em in a fight)
(10:34:27 AM) jeremy: thl: I don't know that a fixed ratio ends up mattering (at all)
(10:34:36 AM) tibbs-cellphone [n=tibbs]  entered the room.
(10:34:41 AM) cweyl: rdieter: :)
(10:34:44 AM) abadger1999: warren: Except FESCo is currently an elected body.  We should get community buyin that it's okay to change that.
(10:34:50 AM) tibbs: Folks, I need to split for a thanksgiving lunch.  I'll see if I canfollow on my phone.
(10:34:51 AM) warren: I think we need to identify the major R&D engineering departments and make sure they have at least ONE representative that is accountable to maintain coordination between FESCo and their department.
(10:34:54 AM) f13: and we might not want them all on 'FESCo' necessarily, however sigs around them sure.
(10:34:54 AM) thl: well, a "at leat 50% of the mebers come from the community" would be a good sign to the community
(10:35:12 AM) f13: and we might actually have a subcommitee that is a release steering commitee that answers to and reports to FESCo
(10:35:23 AM) thl: f13, sounds nice
(10:35:23 AM) warren: abadger1999, absolutely.  RH cannot just streamroll over community decisions.  So this would require ratification.
(10:35:40 AM) warren: thl, I don't think 50% community or setting any number is necessary or a good thing.
(10:35:59 AM) warren: it might look like
(10:35:59 AM) lmacken [i=lewk]  entered the room.
(10:36:25 AM) abadger1999: f13:  I like your proposal
(10:36:34 AM) warren: 1 release engineering, 1 desktop, 1 installer team, 1 java team, 1 kernel team, and all the rest community elected.
(10:36:36 AM) c4chris: f13, sounds good
(10:36:51 AM) f13: the only issue is, I"m a paid employee to do release engineering, I can't necessarily be voted off the island.
(10:37:19 AM) f13: well I could, but that would probabably involve being reassigned or terminated
(10:37:21 AM) warren: yes, thus it makes sense for each department to have a seat on FESCo and be accountable to act as conduit
(10:37:34 AM) jima: f13: that would entail you pissing enough people off.
(10:38:04 AM) cweyl: jima: good point.  all FESCo seats are elected at this point, yes?
(10:38:11 AM) f13: jima: yeah.  I mean, I don't think I"d piss enough people off to ever not be in this position if I wanted to, but would I really need to re-run for election each term?
(10:38:20 AM) thl: well, having "1 release engineering, 1 desktop, 1 installer team, 1 java team, 1 kernel team," sound okay, een if those can't get oted of the island
(10:38:26 AM) jeremy: I think that trying to focus on the governance details might be a little premature.  I'd rather leave that as a "needs to be done".  because in some ways, the board has a bit of a say (which also has community folks)
(10:38:27 AM) abadger1999: warren: Except, there ill be community members that fill those areas as well.
(10:38:31 AM) jima: f13: hmm. tricky point.
(10:38:38 AM) bpepple: jeremy: +1
(10:38:40 AM) thl: but I'd still like to see some sort of ratio "at least foo% from the community"
(10:38:45 AM) jima: cweyl: yes, all are elected thus far.
(10:38:47 AM) thl: jeremy, +1
(10:38:50 AM) f13: jeremy: I agree completely.  I'm just raising some of the concerns.
(10:38:54 AM) jeremy left the room ("Ex-Chat").
(10:39:01 AM) f13: and then he left.
(10:39:02 AM) abadger1999: +1 for putting off decisions on governance
(10:39:03 AM) jeremy [i=katzj]  entered the room.
(10:39:09 AM) warren: Yeah, we don't have to decide anything now.
(10:39:13 AM) f13: jeremy says compiz--
(10:39:26 AM) pygi [n=mario]  entered the room.
(10:39:29 AM) thl: yum install beryl
(10:39:33 AM) ***thl hides
(10:39:35 AM) pygi: f13, me bugging again if you don't mind ^_^
(10:39:47 AM) dgilmore: pygi: after the meeting
(10:39:50 AM) warren: OK, next major topic
(10:39:52 AM) jeremy: thl: doesn't help the fact that the window management aspects are severely crack-laden.  but anyway :)
(10:39:54 AM) f13: pygi: private, there is a meeting going on.
(10:39:59 AM) iWolf left the room (quit: Read error: 110 (Connection timed out)).
(10:40:14 AM) ***pygi is sorry
(10:40:16 AM) thl: warren, okay, so what next?
(10:40:24 AM) warren: lmacken wanted to talk about and demonstrate the next generation fedora update pushing system.  It is amazingly cool.  But he is too busy to talk about it today, so we'll have to do it next week.
(10:41:01 AM) warren: This update system is necessary because *ALL* package updates for a stable release across the entire distro will need update announcements.  Thus we need a low overhead and mostly automated way to do this.
(10:41:31 AM) thl: warren, I put "will community members like it if we switch away from the rolling release model in the future? Is there a alternatie for those that like the rolling release model? We might scare away our contributors otherwise..."
(10:41:32 AM) f13: warren: proposed that they will need 9:
(10:41:34 AM) dgilmore: warren: what ive seen of lmacken's update system  it loks really good
(10:41:53 AM) warren: f13, true
(10:42:02 AM) warren: thl, rolling or not is update for debate
(10:42:05 AM) warren: err
(10:42:07 AM) warren: up for debate
(10:42:10 AM) thl: warren, and we probably need to "merge" the current push scripts somehow with the update stuff from lmacken
(10:42:20 AM) warren: thl, absolutely
(10:42:24 AM) kushal left the room ("Leaving").
(10:42:40 AM) thl: warren, I don't really want rolling
(10:42:49 AM) thl: warren, but some people like it
(10:42:56 AM) warren: I think we want to achieve some kind of balance between rolling and not rolling
(10:43:13 AM) jeremy: it really is a balancing act
(10:43:32 AM) warren: At release time, a "platform" is defined, meaning no version upgrades for gratuitous reasons (just because it is a new version) if it break stuff.
(10:43:39 AM) Belegdol [n=jsikorsk]  entered the room.
(10:43:41 AM) thl: will we still have "updates-testing"? ( I hope so)
(10:43:47 AM) c4chris: I rather like the ideas on
(10:43:48 AM) warren: thl, absolutely yes.
(10:43:48 AM) jeremy: thl: yes
(10:43:57 AM) Belegdol: what is the preferred way to get rid of hardcoded rpaths?
(10:44:05 AM) Belegdol: oops, sorry. meeting
(10:44:06 AM) ***nirik thinks that just rolling devel and only needed updates in released is good.
(10:44:12 AM) thl: well, why don't we rename updates-testing to "updates-foo"
(10:44:19 AM) warren: eh?
(10:44:25 AM) thl: and we push a bit more risky stuff there
(10:44:29 AM) dgilmore: thl: you really do get rolling updates if people update to latest version of foo during a release
(10:44:40 AM) thl: we don#t hae to move it to updates proper later
(10:44:47 AM) dgilmore: weather it is in the way we do it now  or release time snapshot + updates
(10:44:51 AM) jeremy: nirik: the question is how do you define "needed".  I think that realistically, we do a pretty good job today and so things _won't_ drastically change
(10:44:54 AM) dgilmore: the net effect is the same
(10:45:08 AM) jeremy: nirik: the big change is that we want to make it more apparent why updates are occurring for "released" I think
(10:45:19 AM) nirik: well, I think core does ok, extras people push updates all the time just for version updates or whatever.
(10:45:39 AM) bpepple: nirik: agreed.
(10:46:09 AM) warren: and there are many cases where version upgrades in core and extras make sense
(10:46:12 AM) jeremy: nirik: there are lots of version updates for core, too.  although maybe they're more "considered" than "new version, push it!"
(10:46:39 AM) nirik: in core who decides if a update is ok? engr? or release manager? someone else?
(10:46:41 AM) c4chris: warren, +1
(10:46:57 AM) jeremy: nirik: engineer
(10:46:59 AM) warren: So... maybe we have to more explicitly codify exactly what is a good and bad.
(10:47:29 AM) c4chris: the trouble is always in dependancies...
(10:47:36 AM) warren: Also, we need tools to automatically tell us "Uh... your proposed update requires 117 other packages to be rebuilt.  Are you sure?"
(10:47:52 AM) c4chris: right
(10:47:56 AM) thl: warren, +1
(10:48:01 AM) bpepple: warren: +1
(10:48:12 AM) rdieter: "your update requires 117 packages to be rebuilt, what kind of crack are you smoking?" (:
(10:48:30 AM) dgilmore: rdieter: thats a better response
(10:48:37 AM) warren: Thought crime report is then sent to FESCo.
(10:48:37 AM) jeremy: heh
(10:48:43 AM) nirik: perhaps it should still be engr, but for release versions you have to have a bug justifying it... then at least people could see why and have a blocker to track all the in process ones?
(10:49:04 AM) warren: Beyond some threshold of breakage, maybe FESCo would need to approve an update to happen?  Dunno.
(10:49:17 AM) rdieter: nirik: good idea in general, not sure if it needs to be mandated however.
(10:49:18 AM) bpepple: warren: that might be a good idea.
(10:49:19 AM) thl: warren, FESCo? No
(10:49:28 AM) nirik: rdieter: if it
(10:49:31 AM) thl: warren, someone from FESCo could to that
(10:49:33 AM) nirik: 's not, no one will do it. ;)
(10:49:46 AM) warren: thl, someone other than yourself in FESCo? =)
(10:49:55 AM) nirik: updates SIG? :)
(10:50:02 AM) thl: well, a "release manager" could handle that
(10:50:13 AM) warren: thl, I meant, only for bigger breakage updates, needing approval from (someone) to make it happen.
(10:50:13 AM) c4chris: ain't we supposed to have a release engineer ? :-)
(10:50:16 AM) warren: We can't be totally inflexible.
(10:50:17 AM) thl: or some team of "release managers"
(10:50:37 AM) rdieter: nirik: we're already raising the bar from what Extras needs to do updates, I think raising it any higher may be not worth it.
(10:50:47 AM) thl: warren,  bigger updates would need approval sounds okay to me
(10:50:52 AM) warren: Anyway, we don't need to decide things today.  Just put out the ideas.
(10:50:53 AM) cweyl: rdieter: +1.  we're all volunteers
(10:51:02 AM) cweyl: (well.  almost all<grin>)
(10:51:24 AM) thl: warren, where do we discuss all the stuff in the next months?
(10:51:27 AM) thl: fedora-devel?
(10:51:29 AM) warren: I don't recall there being anything else big requiring FESCo approval, although my brain just might be fuzzy.
(10:51:31 AM) thl: fedora-maintainers?
(10:51:34 AM) thl: fedora-extras?
(10:51:41 AM) warren: mmm
(10:51:45 AM) warren: fedora-maintainers?
(10:51:56 AM) rdieter: maintainers, sure (for lack of anything better)
(10:51:59 AM) bpepple: fedora-maintainers: +1
(10:52:07 AM) thl: some people might want to throw opnions in een if they are not (yet) maintainers
(10:52:10 AM) c4chris: fedora-maintainers: +1
(10:52:14 AM) abadger1999: Have e solved getting everyone subscribed to maintainers yet?
(10:52:18 AM) warren: Also, you might want to participate in the fedora-scm temporary SIG to help testing and selection of whatever VCS this new distro will use.
(10:52:24 AM) dgilmore: maintains +1
(10:52:31 AM) abadger1999: Seg was still not on it a few weeks ago.
(10:52:43 AM) dgilmore: with a merge  we should get most of the extras lists closed
(10:52:59 AM) warren: Yeah, but we can figure out those details later.
(10:53:18 AM) bpepple: dgilmore: agreed.  there's way too many lists.
(10:53:32 AM) abadger1999: For those who didn't follow the summit -- switching SCMs was pushed back a bit.
(10:53:37 AM) thl: abadger1999, I'm a admin on -maintainers these days
(10:53:46 AM) warren: more testing and experimentation is needed for SCMs
(10:53:54 AM) thl: abadger1999, I make sure that people that ask to subscribe get subscribed if they are maintainers
(10:54:58 AM) abadger1999: thl: k
(10:54:59 AM) thl: mmm, quite quiet
(10:55:04 AM) thl: warren, what next?
(10:55:33 AM) warren: I don't have any other major topics.
(10:55:52 AM) warren: for FESCo specifically
(10:56:04 AM) thl: should we also have a fedora buildsys sig temporary?
(10:56:12 AM) thl: just wondering
(10:56:17 AM) warren: we alreayd have a buildsys list
(10:56:26 AM) warren: buildsys development will be ongoing forever
(10:56:39 AM) rdieter: I'd say interested parties just sit on buildsys-list and #fedora-buildsys
(10:56:41 AM) warren: fedora-scm is temporary with a deadline in order to make it clear that we need to experiment and make a decision soon
(10:57:07 AM) abadger1999: thl: Should I send out a mail to -extras mentioning that people should subscribe to -maintainers if they're a package maintainer?  Or are we going to automate that at some point?
(10:57:24 AM) thl: abadger1999, please send such a mail
(10:57:31 AM) abadger1999: warren: Are we making that deadline so we have time to implement for FC7 or are we pushing for post-FC7?
(10:57:42 AM) abadger1999: thl: Adding to my list
(10:57:43 AM) rdieter: agadger1999: the former.
(10:57:43 AM) thl: abadger1999, automating that stuff won#t be easy
(10:58:04 AM) warren: abadger1999, I was under the impression that fedora-scm needs to implement something this year.
(10:58:07 AM) thl: abadger1999, we could do that if we really want to start a maintainers-annouce list ( I still think we need that)
(10:58:19 AM) rdieter: abadger1999: actually, no, it'll happen when it happens. (better sooner rather than later).
(10:58:41 AM) rdieter: warren: +1
(10:58:44 AM) jeremy: warren: umm, the timeline isn't this year
(10:58:57 AM) jeremy: there's very little left in this year :)
(10:59:01 AM) warren: true
(10:59:15 AM) warren: jeremy, when do we need the new SCM deployed?  like before the fdcon?
(10:59:16 AM) ***rdieter mind is foggy now, it's all a haze.
(10:59:30 AM) jeremy: warren: no
(10:59:56 AM) jeremy: warren: the scm stuff doesn't block f{e,c,u,,}7
(11:00:11 AM) abadger1999: Just wondering because a straight cvs-dist => $SCM-dist is easy to imagine and get buyin to do.  But designing alternatives has _a_lot_ of room for experimentation.
(11:00:18 AM) warren: Early  February we are proposing as a fedora developer conference, somewhere in or near Boston.  We will fly in top fedora developers and put them into hotels.  And we focus on hacking on the distro and tools.  Probably much of that will be the mass review.
(11:00:46 AM) warren: jeremy, I mean, the decision could just be deploy cvs as-is
(11:01:02 AM) warren: jeremy, or do you mean we just import everything into /cvs/extras and rename it?
(11:01:20 AM) jeremy: abadger1999: yeah, but we don't want to have to make people learn something new in 3 months and then something else in 9
(11:01:34 AM) abadger1999: jeremy: Exactly.
(11:01:36 AM) jeremy: warren: import everything into /cvs/extras and probably add a symlink was the proposal
(11:01:42 AM) warren: jeremy, ah
(11:01:47 AM) jeremy: abadger1999: so I'd much rather have us take the time to fully investigate stuff than rush into a new scm
(11:01:53 AM) warren: anyway, if you're interested in the SCM stuff, please join that group
(11:02:02 AM) warren: jeremy, when will the fedora-scm stuff be created?
(11:02:29 AM) abadger1999: Yeah.  Which is why making it a post-FC7 goal might be better.
(11:02:31 AM) jeremy: warren: hopefully in the next few days.  because I want the thread on fedora-maintainers to stop :P
(11:02:37 AM) jeremy: abadger1999: that's how I'm personally seeing it
(11:02:39 AM) dgilmore: anyone intretsed in any of the buildsys scm stuff please join infrastructure team
(11:03:13 AM) warren: any other topics?
(11:03:32 AM) thl: I don't think so...
(11:03:59 AM) thl: we really should try to sort out the governance model quickly
(11:04:21 AM) c4chris: thl, yup, that'd be nice
(11:04:25 AM) warren: what aspect of governance specifically?
(11:04:31 AM) thl: FESCo future
(11:04:43 AM) thl: how many members
(11:04:53 AM) thl: re-vote the exsiting members
(11:04:55 AM) warren: one thing that bothered me about past discourse is the desire to shrink the size of FESCo.
(11:05:00 AM) warren: if everyone came to meetings I would agree
(11:05:07 AM) warren: but usually very few show up
(11:05:28 AM) warren: I think a larger FESCo works fine and never was a problem.
(11:05:35 AM) bpepple: warren: agreed.
(11:05:39 AM) rdieter: warren: +1
(11:05:51 AM) thl: well, when we had a larger fesco een less members showed up ;)
(11:06:03 AM) Belegdol left the room (quit: "Leaving").
(11:06:05 AM) thl: but yes, I think we should have a bigger FESCo soon
(11:06:09 AM) abadger1999: warren: The only difficulty with a larger fesco is that it becomes impossible to vote against someone.
(11:06:11 AM) rdieter: From seeing how a smaller packaging comittee works (which has been painful at times), bigger sometimes is better.
(11:06:18 AM) warren: abadger1999, how so?
(11:06:26 AM) dgilmore: thl: yeah  we should  we will have a much larger job
(11:06:52 AM) warren: larger fesco as an umbrella organization, and sub-commitees can split from it to focus on specific problems and come back with recommendations?
(11:06:53 AM) abadger1999: If there's ten seats and twenty people you have a good chance to cast votes that can keep someone you don't like from getting a seat
(11:07:00 AM) abadger1999: not so if there's 19 seats.
(11:07:00 AM) thl: warren, +1
(11:07:07 AM) Rathann [n=rathann]  entered the room.
(11:07:34 AM) rdieter: abadger1999: so what you really want are anti-votes? (:
(11:07:46 AM) warren: Yes, we want anti-votes
(11:08:03 AM) warren: Oh, here's another thought.
(11:08:14 AM) abadger1999: Yeah -- maybe we should revisit the elections policy and have it based on getting a certain minimum number of votes or something.
(11:08:16 AM) warren: The new shiny Fedora distro would need a ranking system.
(11:08:22 AM) warren: I thought about it something like this:
(11:08:31 AM) warren: FD0 (currently doesn't exist)
(11:08:35 AM) warren: FD1 (cvsextras)
(11:08:39 AM) warren: FD2 (something in between)
(11:08:41 AM) warren: FD3 (sponsor)
(11:08:44 AM) warren: FD4 (something in between)
(11:08:52 AM) warren: FD5 (FESCo current or emeritis)
(11:09:09 AM) warren: You are only eligible to run for FESCo if you are FD3 or higher.
(11:09:16 AM) c4chris: contributors, you mean ?
(11:09:20 AM) thl: warren, I'd like to hae cvsextras and buildsys
(11:09:26 AM) nirik: FD2 == cvsbugs? (ie, can review/approve packages for already existing sponsored people)
(11:09:36 AM) thl: warren, to give people access to commit stuff (e.g. upstream maintainers)
(11:09:40 AM) thl: but bot to build it
(11:09:48 AM) warren: We can figure out the exact meaning of the ranks later.
(11:10:27 AM) warren: There would be a parallel rank structure for RH engineers.  For example RD3 would have the same group ACL access as FD3, but cannot sponsor FD1 membership.
(11:10:37 AM) warren: I'm trying to put together a proposed map of all this.
(11:10:50 AM) warren: I'll post to list when I have a strawman.
(11:11:04 AM) warren: Basic idea is higher rank *somehow* is able to promote people to the previous rank(s)
(11:11:15 AM) warren: And we'll need some way to demote people too =)
(11:11:37 AM) warren: "We made a clear mistake in upgrading you to FD1..."
(11:11:44 AM) rdieter: start hoarding all your fedora charma/experience points...
(11:11:49 AM) thl: warren, btw, I created a page in the wiki for that effort:
(11:11:58 AM) warren: thl, cool
(11:12:03 AM) thl: warren, you you make sure that it maintained please?
(11:12:17 AM) thl: to allow other FESCo members to track progress?
(11:12:23 AM) warren: My strawman for a ranking system would be related to, but not exactly tied to the proposed experience point system.
(11:12:33 AM) thl: and to get up2date if they were on vacation?
(11:12:43 AM) warren: thl, k
(11:12:57 AM) thl: proposed experience point system ?
(11:12:59 AM) warren: thl, I totally see the alternate paths as ways of gaining ranks.
(11:13:07 AM) pygi left the room (quit: Nick collision from services.).
(11:13:18 AM) thl: warren, sure (and I like that idea)
(11:13:18 AM) warren: thl, that's somewhat related, I have to write up about that separately.
(11:13:27 AM) rdieter: warren: it's an excellent notion.
(11:13:42 AM) bpepple: warren: Sounds like a good plan.
(11:13:57 AM) warren: I see ranks in general as broad group-based ACL's, permissions of what they can do, etc.
(11:14:14 AM) c4chris: warren, is a level 17 human packager with herding skills ... :-)
(11:14:18 AM) warren: There would be different scales of ranks... FD (developer), FI (infrastructure), RD (rh engineer) etc.
(11:14:28 AM) warren: c4chris, with a +3 sword
(11:14:33 AM) rdieter: warren: it certainly would make acl's easier.
(11:14:37 AM) warren: rdieter, yep
(11:14:42 AM) c4chris: warren, exactly :-)
(11:15:01 AM) thl: who will design the acls for cvs?
(11:15:10 AM) warren: thl, we have to figure that out later
(11:15:25 AM) ***rdieter is still keeping his flame-resistent +4 shield of righeousness
(11:15:32 AM) thl: just one note: I'd like to see co.maintainership for pacakges that come from core, too
(11:15:40 AM) thl: at least in the longer term
(11:15:43 AM) rdieter: thl: absolutely, +1
(11:15:48 AM) warren: thl, yes, that is totally the thinking.
(11:15:59 AM) ***rdieter just had a chat about that with a rh engineer this morning...
(11:16:13 AM) warren: thl, that both ranks allow people access, and core maintainers delegate access either to entire groups or specific people.
(11:16:15 AM) thl: and somehow aoid the stupids fights like the one with firefox on fedora-deel in the last week
(11:16:26 AM) warren: Most of core would be wide open for FD4+, for example.
(11:16:32 AM) thl: the maintainer and his opinion is important
(11:16:49 AM) thl: but sometimes doing something that is good for the porject is more important
(11:16:50 AM) warren: the multilib one?
(11:16:56 AM) thl: warren, yes
(11:17:05 AM) ***rdieter is glad he missed that thread.
(11:17:05 AM) warren: That fight isn't over.  Good thing I have a +3 sword.
(11:17:16 AM) ***c4chris needs to go hunt food rations (and other house chores) shortly...
(11:17:26 AM) thl: yeah, we are getting later
(11:17:28 AM) thl: late
(11:17:30 AM) warren: I think we should end this meeting?
(11:17:37 AM) thl: anything else we need to discuss today?
(11:17:42 AM) nirik: I have one quick item to bring up...
(11:17:47 AM) thl has changed the topic to: FESCO meeting -- free discussion
(11:17:50 AM) thl: nirik ?
(11:18:20 AM) nirik: Jonathan Steffan (daMaestro) has requested sponsorhip.. he wants to take over the orphaned zope/plone packages.
(11:18:30 AM) nirik: he apparently maintains them for fedora unity?
(11:18:50 AM) nirik: Nov 15 12:15:03 <daMaestro>     it would just be easier for me to take over the packages s
(11:18:50 AM) nirik: upeeing as how i end up building fixes for fedora unity as we move along with plone.. and
(11:18:51 AM) nirik: then wait on extras to catch
(11:19:13 AM) c4chris: nirik, are you a sponsor ?
(11:19:16 AM) warren: Does he seem to have an understanding of fedora processes?
(11:19:20 AM) nirik: he has no other packages in review... I don't know anything about him.
(11:19:32 AM) nirik: he said some of the other sponsors might know him and his work...
(11:20:06 AM) thl: I know him a bit, but I don't know how good his packaging skills are
(11:20:08 AM) ***bpepple isn't familiar with him.
(11:20:08 AM) nirik: c4chris: yes, but I don't want to sponsor him, since I know nothing about him. I thought I would bring it up in case anyone here did...
(11:20:19 AM) dgilmore: i only know of the work he has done in the infrastructure space
(11:20:20 AM) thl: nirik, maybe we should ask aurelien again (the old owner iirc)
(11:20:23 AM) c4chris: nirik, k, understood.
(11:20:35 AM) warren: I propose just giving him FD1 (cvsextras) if he agrees to touch only the packages that he owns, follow the fedora packaging guidelines and FESCo decisions.
(11:20:38 AM) thl: nirik, he could sponsor him and watch him in the beginning
(11:20:40 AM) drpixel [n=drpixel]  entered the room.
(11:20:43 AM) warren: I mean, how much damage can he do to his own packages?
(11:20:44 AM) abadger1999: This ould be ideal comaintainer territory.
(11:21:01 AM) bpepple: abadger1999: agreed.
(11:21:02 AM) thl: abadger1999, agreed
(11:21:08 AM) nirik: yeah, this would be great for real co-maintainerness. :)
(11:21:10 AM) warren: BTW, cases like this would be the purpose of FD0 in the future.  not wide access to Extras, but rather only certain packages.
(11:21:28 AM) c4chris: warren, yup, that'll be nice
(11:21:35 AM) warren: let them demonstrate their skills as FD0 as another path toward FD1
(11:21:45 AM) abadger1999: warren:  That would be great
(11:21:52 AM) warren: we don't have the ranking system yet, but I'd say let him prove it with those packages now.
(11:22:03 AM) warren: As long as someone is willing to keep an eye on him.
(11:22:09 AM) bpepple: warren: +1
(11:22:19 AM) rdieter: warren: +1 (my eyes are all spoken for already though)
(11:22:27 AM) thl: I'll ask aurelien (he is the old zope owner, isn't he?)
(11:22:35 AM) thl: and he's a sponsor, too (iirc)
(11:22:46 AM) c4chris: thl, sounds good
(11:22:50 AM) warren: +1 as long as someone makes sure he does the right thing
(11:22:51 AM) nirik: thl: yeah, asking him might be best... since I know not much about zope/plone other than I dislike installing hem. ;)
(11:23:04 AM) thl: k, settled
(11:23:06 AM) rdieter: maybe even ask DeMaestro to see if he can find himself a sponsor?
(11:23:40 AM) thl: rdieter, I'll ask him if he knows any of the sponsors
(11:23:51 AM) thl: anything else?
(11:24:04 AM) warren: Adjourn please.
(11:24:05 AM) nirik: thl: thanks. Would be good to have zope/plone maintained again...
(11:24:26 AM) ***thl will close the meeting in 30
(11:24:42 AM) nirik: FYI, the FE-NEW queue is the lowest I have seen it in a long time. Thanks to all the work of reviewers! :)
(11:24:46 AM) ***thl will close the meeting in 15
(11:25:01 AM) thl: -- MARK -- Meeting end