From Fedora Project Wiki

2007 April 12 FESCo Meeting



  • Brian Pepple (bpepple)
  • Jason Tibbitts (tibbs)
  • Christian Iseli (c4chris)
  • Rex Dieter (rdieter)
  • Toshio Kuratomi (abadger1999)
  • Kevin Fenzi (nirik)
  • Dennis Gilmore (dgilmore)
  • Josh Boyer (jwb)
  • Jeremy Katz (jeremy)
  • Jesse Keating (f13)
  • Bill Nottingham (notting)


  • Tom Callaway (spot)
  • Warren Togami (warren)


Packaging Committee Report

Renaming cvsextras

  • FESCo approved warren's proposal to rename the cvsextras group.


  • f13 discussed the plans for the switch from plague to Koji in Extras.


  • FESCo voted against the plan to delete everything and then do a mass-rebuild for EPEL5, instead of bumping the spec and rebuilding.

Package Conflicts

  • bpepple received from Michael Schwent the tool to identify packages with conflicts, but he hasn't had time to look at it.


* jeremy is here
<bpepple> FESCo meeting ping -- abadger1999, bpepple, c4chris, dgilmore, f13, jeremy, jwb, notting, rdieter, spot, nirik, tibbs, warren
hi everyone.
* jwb is here
<bpepple> who's around.
<tibbs|h> I'm having some home repair done so I may be in and out.
* nirik is here.
thl is on the rabble seats
abadger1999 here
<rdieter> here
<RajeshPandey> here
--- bpepple has changed the topic to: FESCo meeting -- Packaging Committee Report -- spot, abadger1999, rdieter, tibbs, scop
<bpepple> tibbs|h: you want to take this?
<tibbs|h> Two proposals this week.
Yes, I'll run it.
First is a basic statement of the responsibilities of reviewers and packagers during the review process.
If only the wiki were working....
<bpepple> d'oh!
<tibbs|h> This is going to be difficult otherwise, I guess.
<jeremy> indeed
* dgilmore is here
<tibbs|h> Perhaps abadger1999 has the text laying around somewhere.
<abadger1999> Google cache
* c4chris is here
<tibbs> abadger1999: Not quite.
Oh, there it is at the bottom.
Can everyone see that?
<bpepple> yup.
<tibbs> This is intended to replace the first paragraph of the ReviewProcess document.
<bpepple> I don't see any problem with it. +1
<c4chris> +1
<tibbs> abadger1999: Is that the final text we chose?  The cache is from before the meeting.
<jeremy> looks okay I guess
<nirik> +1 from here...
<abadger1999> If it's before the meeting then it's old.
<rdieter> final draft was a worded a little diferent, but the intent is/was the same.
<notting> sorry i'm late. url?
<abadger1999> I can pull the texts directly from the server if you want.
<f13> well, I +1'd it in the packaging meeting, so I'll +1 it here.
notting: wiki go boom, can't look directly at it
<tibbs> abadger1999: That might be better; another alternative would be to postpone, but it would be really nice to get to the conflicts document.
Since folks have been waiting on that.
<abadger1999> What's the URL for the other docs?
If that's what you're asking.
<abadger1999> K.  Let's postpone this portion and I'll pull the text so we can do this later in the meeting.
<bpepple> abadger1999: ok.
<tibbs> OK.  bpepple, let's move on.
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- renaming cvsextras group -- warren
* f13 runs to grab his lunch
<bpepple> jeremy: warren's gone, but he mentioned on the mailing list that he wanted to implement this.
possibly after the meeting, providing we approved the idea.
<jeremy> bpepple: yeah.  and as long as we're not switching the rest of the world at the same time, I'm okay with it
<jwb> fine with me
<bpepple> quick vote then?
<f13> +1
<jwb> the wiki needs updating at the same time though
<c4chris> +1
<bpepple> +1 here also.
<notting> +1
<jeremy> jwb: yes.  warren had a list of things needing updating
+1 even
<jwb> +1
<abadger1999> +1
<nirik> +1 (also if the script that sends out the sponsorship needed needs updating it should be at the same time)
<rdieter> +1
<tibbs> +1
<bpepple> ok, so this is approved.
I'll send a note to warren.
--- bpepple has changed the topic to: FESCO-Meeting -- MISC - koji - f13
<bpepple> f13: you wanted to talk about kojii
<dgilmore> bpepple: we should switch Extras from plague to koji in the next week
<f13> yeah, I sent mail to maintainer's regarding using Koji
outlined what I think needs to happen
<tibbs> Yes, that seems reasonable.
<f13> but it all seems doable.
<tibbs> One question, though:
<f13> it's not hte merger yet, but will ease us into it.
<tibbs> if dist-fc7 inherits from dist-fc6, how do we drop packages?
<f13> dgilmore: we still need an rpm fix for devel.
<nirik> sounds good to me, there should be a clear announcement about any outages and when and perhaps a "about to start" and "finished" emails to maintainers?
<tibbs> Or am I confused about what "inherits" means?
* bpepple hasn't had a chance to read his e-mail today.
<f13> tibbs: you can block a package from dist-fc7
<jeremy> tibbs: there's a method of blocking packages to drop them
<dgilmore> f13: yes we have the srpm and jeremy looked at it
<f13> tibbs: it won't get inherited.
<dgilmore> nasrat has not shown up
<tibbs> OK, no problem.
<jwb> f13, is core still using brew?
<f13> jwb: yes
<dgilmore> jeremy: can you get the patch into rpm
<notting> f13: any progress on the build capacity tests?
<f13> we don't have enough build power in the colo to switch over core yet.
<tibbs> I'm looking forward to some "how to do cool things with koji" document.
<f13> notting: yeah, doesn't look horrible, will be pretty good with a couple more boxes.
<jwb> f13, can the core build machines be switch to koji though?
<f13> notting: crap, I knew I forgot to cc somebody on those tests.
jwb: no.
<notting> f13: buy buy buy!
<jwb> f13, because of RHEL?
<dgilmore> jwb: how did you go ppc wise?
<f13> jwb: because of SarBox and essentially we'd be granting access to people outside of Red Hat to machines inside the firewall.
<jeremy> dgilmore: will try to get to that this afternoon.  a little busy with summit live dvd finagling
<dgilmore> jeremy: :)  thanks and understood
<notting> f13: mmcgrath: boxes on order?
<jwb> f13, no. i don't mean merge yet.  just run separate koji builder on the existing core machines
<f13> notting: I've been granted a virtual blank check to get it done, we just have to figure out what is actually necessary for now until we can deploy the blade center.
<jwb> dgilmore, Quad G5 mac is available whenever mmcgrath wants it
<f13> jwb: I'm not following.
jwb: what good would that do?
<jwb> f13, do it there first where it's controlled and contained before releasing it to extras?
<f13> jwb: lets talk about that after the meeting k?
<tibbs> Obviously we just need a few racks of populated blade centers.
<dgilmore> jwb: can you help us price up som 1U power5 boxes
<jwb> dgilmore, i don't think IBM makes 1U power5 boxes
<dgilmore> jwb, f13:  lets talk hardware with mmcgrath after the meeting
<f13> so, is there anybody that is in disagreement with going forward?
dgilmore: yeah, lots of emails are flying by
<jwb> f13, i'm asking why extras first, that's all...
<f13> jwb: lighter load, existing builders, easy to move.
<bpepple> f13: I don't have any problem with it.
<c4chris> I'm fine with the move
<tibbs> As long as it's relatively painless for the maintainers, then I think the sooner the better.
<f13> jwb: Core is essentially using the same codebase, just a different instance of it.
<notting> f13: don't have enough info to ask more questions - behind on mail.
<jwb> f13, ok
was just trying to get some feel for how much it's been tested
seems fine with me
<jeremy> jwb: due to.. ummm... "similarity" with other things, the underlying bits of koji are quite well tested.
<nirik> yeah, on-ward to f7 victory. ;)
<jwb> jeremy, that's great then
<f13> which really, is mock
<jeremy> nirik: indeed!
<nirik> when were you guys thinking of doing the changeover?
<f13> nirik: that really depends on how soon we can get the needed software bits written
<dgilmore> nirik: when we can
<f13> nirik: now that we have fesco buy-in, we can make that something of a priority
* f13 had hoped mschendent would have responded by now...
<nirik> ok, sounds good. I would like to see updates announced to maintainers or something to keep people in the loop... ;)
<f13> mschwendent?
<dgilmore> f13: he resonds on some things and not on others
<f13> nirik: yep, I"ll constantly drop notes there
<tibbs> Did he have concerns he voiced earlier?
<f13> not that I"m aware of
<bpepple> f13: anything else in regard to koji?
<f13> I can't tell if he's even noticed or not
<dgilmore> tibbs: no we need to tie the existing push scripts into koji
<thl> will EPEL switch to koji at the same times as Extras?
<dgilmore> thl: yes
<thl> dgilmore, k, was just wondering
<dgilmore> thl: well proberlly
<thl> befcause f13's mail had a "(devel?)" in it
<f13> thl: I can't say that right now.
<nirik> dgilmore: so will the 5 rebuild take place now in plague or in koji? I guess it depends on timeframe..?
<f13> I'm planning just devel/ for now
<dgilmore> nirik: depends on timing
<f13> as to not disrupt released product streams
<tibbs> Does it look to be difficult to fix up the push scripts?
<f13> tibbs: in theory, no.  BUt i Haven't actually looked at the scripts
<nirik> so fc5/fc6 builds will be in plague still, and devel will use koji?
<f13> nirik: that's my initial plan.
<dgilmore> nirik: and make build will do the right thing
<f13> once devel use is solid, we can migrate fc6/5 too
<nirik> you should make sure to check the case of 'make plague' (which I sometimes use... )
<f13> sure, balk if on devel/
<nirik> yeah.
<notting> erk. we never did announce anything publicly about fc5, did we?
<dgilmore> notting: no
<f13> notting: nope!
<jwb> f13, why?
<f13> jwb: because plague buildroots won't be updated for devel builds.
<dgilmore> notting: in theory FC-5 will stop support when Test4 is out
<f13> jwb: and push scripts migh tnot pick up anything build for devel/ in plague
<notting> dgilmore: ?
<thl> dgilmore, test4 or test3?
<notting> dgilmore: pretty sure the 'new' thing was a month or two post f7
<f13> thl: test3 is out already.
notting: we were waiting for RH buy in on that.
<thl> f13, ohh really? ;-)
<jwb> f13, sorry, i meant why not switch to koji across the board?
<dgilmore> notting: that was agreed upon for future releases  but never backdated
<tibbs> Due to lack of clarity, I'd wager that many folks are expecting the new thing to apply to FC5 as well as FC6.
<f13> jwb: I don't want to risk an interruption in delivering updates to FC6
<bpepple> tibbs: agreed.
<jwb> you mean FE6
<f13> jwb: right, both because we weren't planning on merging Core 6 either.
not right away
<EvilBob> dgilmore: someone at some point made the change retroactive for the "active" releases
<f13> if we break rawhide, well it's rawhide.  I don't want to break a live product, especially if there is a security issue that comes up.
<jwb> ok.  i find it a bit confusing that two buildsystems will be in use, but i guess there is some pain like that to be expected
<thl> EvilBob, sure? I never saw that annouced anywhere, but maybe I missed it
<dgilmore> EvilBob: that was not communicated anywhere
<jeremy> jwb: 3 build systems for some of us.  whee! :)
<f13> dgilmore: I think we talked about it at a Fedora Board meeting
<jwb> heh :)
* thl wonders if mspevack is around to clarify
<EvilBob> IIRC it was in a board meeting
<f13> a discussion of how retroactive to make the new lifespan.
<dgilmore> well it needs to get out in the wild if that is the case
anyway lest move on
<thl> dgilmore, +1
<bpepple> dgilmore: agreed.
* rdieter recalls the "retroactive" discussions as well.
<nirik> I think that was waiting on RH buyin for doing security updates for fc5 longer.
<thl> rdieter, discussions yes, but no final decisions iirc
<rdieter> notting said no one yelled, so we called it good. :)
<thl> :-)
<bpepple> ok, we should probably move on.
<f13> dgilmore: we wanted to get RH buy in before trying to get fesco buy in.
<dgilmore> f13: thats where i thought things were at
waiting on RH
<notting> hey, the wiki's alive
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- Package Database - abadger1999
<f13> dgilmore: that's what we've just spent 10 minutes confirming.
<abadger1999> bpepple: Nothing to report.  Move on.
--- bpepple has changed the topic to: FESCo meeting -- Packaging Committee Report -- spot, abadger1999, rdieter, tibbs, scop
<f13> dgilmore: that it's still waiting on RH
<bpepple> since the wiki's back up.
tibbs thl tdiehl
<f13> now that the wiki is up... (:
<bpepple> tibbs: want to go back to this?
<tibbs> I'm still getting "down for maintenance"
<jeremy> it is?
<notting> ...and it's down again
* f13 shakes his fist at mod_python
<bpepple> d'oh! alright moving on again....
<tibbs> abadger1999: Any luck with extracting the texts?
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- Package Conflicts
<tibbs> Ah, there's the conflicts draft.
<bpepple> I talked to mschwent about this, and got his tools.  He didn't sound interested in heading it, though.
<tibbs> So, we have enough to proceed with the PC report, and the conflicts draft ties in.
--- bpepple has changed the topic to: FESCo meeting -- Packaging Committee Report -- spot, abadger1999, rdieter, tibbs, scop
<tibbs> First off is
<notting> 2
<tibbs> Not much there really, but questions cropped up recently about this so we felt a bit of clarification was in order.
<c4chris> looks fine +1
<bpepple> +1
<notting> +1
<f13> +1
<tibbs> +1
<nirik> +1
<tibbs> Anyone else?
<jwb> +1
<rdieter> +1
<tibbs> OK, next is
<f13> +1 on that too
<nirik> +1
<tibbs> A bit more meat here; I hope everyone had a chance to read over it.
<c4chris> +1
<jwb> reading
<rdieter> +1
<jeremy> looks pretty reasonable.  +1
<tibbs> It is expected that we'll find other situations that will need to be clarified, so we'll add more sections to this as necessary.
<f13> tibbs: did the PC approve the Conflicts draft?  I don't remember it
<bpepple> +1
<f13> but then again....
<tibbs> f13: Yes, voting was 5-0
<abadger1999> On Tuesday.
<jwb> +1
<tibbs> And there's a +1 from you further down in the logs which I wasn't sure about.
<notting> +1
<f13> tibbs: oh right, yeah I +1'd it
sorry, this week has been... fun.
<tibbs> f13: I'll update the minutes to indicate that.
People have been wanting this conflicts draft for some time now, so it's good to get motion on it.
BTW, +1
<notting> tibbs: i think the +1 was implied in your previous statement
<tibbs> Anyone else?
Remember to get your comments out there so that we can consider them before Tuesday when we'll write this up.
<tibbs> bpepple: Assuming, of course, you indicate that this has passed.
<tibbs> Otherwise that's all from the PC.
<bpepple> tibbs: I consider these guidelines approved by FESCo.
<tibbs> abadger1999: Thanks for extracting that text.
<abadger1999> tibbs: No problem
<bpepple> tibbs: anything else, or should we move on?
<notting> tibbs: a 'what to do if upstream refuses to rename' section might be good. not that i have any ideas in that regard
<f13> I think that's it.
<tibbs> Yes, that's it from us.
<bpepple> ok, moving on....
--- bpepple has changed the topic to: FESCO-Meeting -- EPEL
<bpepple> anything in regard to EPEL need to be discussed?
<jwb> yes
* thl send some notes from this week to the list
<tibbs> notting: If you do have any ideas about that, please let us know.
<thl> one hour or so ago
<jwb> there was the buildroot issue Axel wanted acked by FESCo
<notting> tibbs: get out baseball bats and beat upstream? :)
* bpepple hasn't had a chance to read his e-mail today. :(
<thl> bpepple, it's about the "mass-rebuild of EPEL5 now that we soon have RHEL5 final on the builders"
bpepple, it was voted to delete everything and just rebuild
<bpepple> thl: ok.
<jwb> -1
<thl> everything, without chaning ENVR
<f13> er..
<jwb> yeah, -1
<f13> that may not bode well for clients whom already ahve stuff installed
<nirik> (note: only EPEL-5... not 4)
<thl> f13, tell those that voted like that
<f13> as noted many times before, packages changing checksums and such get messy
<jwb> f13, it was noted in their discussion.  apparently it didn't seem that big of a deal
* thl disliked that plan, too
<jwb> are we voting on this yet?
<f13> *shrug*  I don't run rhel5 so I won't get effected by it.
<bpepple> jwb: Yeah, we should do a quick vote.
<jwb> f13, more than rhel5
<tibbs> I'm still not understanding why you wouldn't want to bump, and I read the IRC logs.
<jwb> tibbs, me either
<dgilmore> tibbs: becaue people did not want to fork the spec
<jwb> that is just lazy
<thl> jwb, +1
<dgilmore> i wanted to add a .1 and rebuild
<jwb> you're pissing on your users because you don't want to make a 2 character change
<tibbs> Ah, that is a point, but I don't think it's a terribly good point.
<jwb> dgilmore, that would be very acceptable
<c4chris> yea, .1 and rebuild
<rdieter> dgilmore: +1
<tibbs> The spec will diverge pretty much immediately anyway.
<jwb> right
<thl> dgilmore, why did you vote for deleting the packges then?
* thl is confused
<notting> ? you don't need to fork the spec. just b/c the release changes, doesn't mean you have to build and push for older releases
<jwb> notting, fork it vs. the fedora spec
<f13> notting: er, they have to bump the spec there, but nowhere else, so now the specs are diverged
<notting> *horrors*
<tibbs> As I understand things, EPEL has no reason to attempt to keep any kind of release ordering with Fedora.
<dgilmore> thl: i was confused by then.
<f13> not that I find anything _wrong_ with that.
<tibbs> So it's not even appending ".1"; just bump the release.
<f13> nod
<thl> dgilmore, np, I was just confused now
<notting> thl: well, two issues. i'd be all for 'rebuild and delete all old packages', but with a release bump
<thl> tibbs, some people prefer to appending ".1" ovefr bumpin the release
<f13> is there a call for fesco vote?
<thl> I think they have a point
<jwb> f13, axel requested one
<f13> or a point 1
<thl> notting, sounds fine for me
<c4chris> :-)
<tibbs> OTOH, not rebuilding at all seems to be working for Fedora at this point.  What's the reason they absolutely must be rebuilt?
<abadger1999> tibbs: If they want to use the vanilla spec later, using .1 lets them come back on the next Fedora Release rather than the next upstream bump
<nirik> tibbs: they were build against beta1
<tibbs> abadger1999: Extremely good point.
<f13> abadger1999: but that actually overwrites history
unless they merge that .1 somewhere into the history of hte FEdora spec
<nirik> abadger1999: yeah, changelog is lost then if you merge
<tibbs> nirik: And we have .fc6 packages in F7; surely F7 diverges from FC6 more than rhel5b1 diverges from rhel5release.
<thl> f13, is that really a big problem if it was just a "rebuild" in the chanelog?
<notting> dgilmore, this is only rebuilding things actually built for EPEL, not everything in EPEL cvs, right?
<thl> notting, yes, only what has been build up to now
<f13> thl: it's not a really big problem, but I generally don't like to see history get stomped
<nirik> tibbs: yeah, you would think so... dunno for sure.
<f13> and who k nows what happens with the rebuild, something may end up needing changed to build again against RHEL5 GA
<thl> f13, agreed; I think in this case it's still not nice, but acceptable
<f13> tibbs: you'd be surprised what all changed from B1 to GOLD
* rdieter thinks we're not here to (re)make epel's decision for them (or not?), just ack or nack it.
<f13> -1
(for their current plan)
<jwb> -1
<c4chris> (plan == rebuild and no bump, right)
<jwb> rdieter, but we can nack with a suggested improvement
<thl> c4chris, yes
<bpepple> c4chris: correct.
<f13> c4chris: yep
<c4chris> k, -1 then
<notting> -1
<bpepple> -1 here also.
<tibbs> Yeah, I hate to be an obstruction, but -1 to rebuilding with no bump.
<abadger1999> -1
<thl> jwb, I can take care of that if you want; i was against this in any case ;-)
<bpepple> so it looks like we against EPEL suggested plan.
<jwb> thl, great
<thl> bpepple, I'll get that out to epel and will take care of it
<dgilmore> notting: yeah just whats built
<bpepple> thl: great, thanks.
<thl> but I'd like to discuss something else:
do you guys here really want to ACK or NACK each decission from the packaging commitee or the epel steering committee?
wouldn#t it be easier to discuss it on the list
<bpepple> thl: personally I don't.
* nirik voted for just wipe and rebuild, but doesn't really care that much... if everyone thinks there needs to be a bump we can do that.
<thl> and bring only controversial topics (like the last one that just got nacked) up here?
that might be a lot quicker
and easier for everyone
that how it was planed and done with the PC months ago, too
<f13> except what's controversial for FESCo may not seem controversial to PC or EPEL sigs
<jwb> thl, it would be helpful for a set amount of time to pass by before actually moving forward with the decision
<thl> f13, that why the summaries get posted to the list
<jwb> so FESCo has time to review and speak up
<thl> jwb, sure
<tibbs> The PC stuff can go either way, but it seems that some things just don't generate comments unless there's actually a vote called.
<abadger1999> I'd like to see a report of issues discussed/decided on.  But not necessarily discussion or ACK/NACK.
<bpepple> abadger1999: +1
<jwb> abadger1999, +1
<thl> jwb, the rule was that the meeting minutes had to be send to the proper place at least 24 hours before the next fesco meeting
abadger1999, +1
the question is: where to send it
<jwb> thl, right, but i mean if we're not going to ack/nack everything, wait a week or so before actually implementing the decisions
<thl> can I send it to the FESCo list directly?
or fedora-maintainers?
jwb, we are in agreement
<notting> maybe just give fesco a week of veto power?
<bpepple> I would lean towards either maintainers or devel.
<f13> -maintainers
<jwb> notting, right
<thl> bpepple, devel is public, so might be better
<f13> notting: I can agree to that.
<jwb> -maintainers
<tibbs> I admit I haven't always been able to get the PC summaries out a full 24 hours previous so far I've made at least the day before.
<thl> why exclude the rest of the world
<f13> or even -maintainers-announce
thl: because the rest of the world doesn't have veto power.
<thl> tibbs, well, then the veto power extends by one week
f13, but the rest of the world should not be excluded
<f13> no reason why maintainer's couldn't be read-only
<thl> anyway, let's stop here for now
we can work out the details on the list
<f13> -devel is horribly noisy to try and run decisions through
<thl> that might be easier
<bpepple> thl: thanks. moving on....
--- bpepple has changed the topic to: FESCo meeting -- Free discussion around Fedora
<nirik> is lmacken around to give a quick status update on bodhi?
<jwb> what is bodhi
* bpepple hasn't heard of it either.
<thl> jwb, the new updates system from lmacken
<wwoods> the updates system
<jwb> huh?
<notting> jwb: patrick swayze in point break?
<jwb> for update announcements?
notting, nice :)
guess lmacken isn't around? Oh well. Just a thought.
<wwoods> for getting new builds into updates-testing and then pushing them to updates
<jwb> f13, oh... does koji do scratch builds?
<f13> yes
<jwb> and they'll be enabled for extras?
<f13> multiple levels of scratch builds.
* nirik cheers
<f13> you can A) build with --scratch, which doesn't tag the build, and puts it in temporary storage that will get flushed.
<jwb> cool
<f13> B) build with --no-tag which will import the build into the database, but not actually tag it, but you can tag it  later to be included.
* thl send something to the list for FESCo discussion (  ) , but that was probably to late to discuss it today
<f13> or C) build for a -candidate tag so you can test it and later "move" it to a live tag.
oh and --scratch builds don't get imported into the DB and thus can never be tagged for any collection
<jwb> f13, there is a doc that will be sent out on how to use koji right?
<rdieter> no, you have to figure it out yourself, no cheating and asking for help either. :)
<f13> a simple one to start with
warren is supposed to be working on a FAQ
<bpepple> thl: I was thinking about bringing that up at next week's meeting, since I hadn't heard back from Max about it.
<jwb> rdieter, then you're screwed.  you can't even make simple CVS requests correctly
<thl> bpepple, k
* jwb runs
<thl> bpepple, maybe discussing it first on the list is easier anyway
<rdieter> yep. :)
<jwb> f13, ok great
<bpepple> not to mention we're running late. ;)
<jwb> :)
<bpepple> anything else people want to discuss before wrapping up the meeting?
<c4chris> nothing here
* bpepple will end the meeting in 60
bpepple will end the meeting in 30
bpepple will end the meeting in 15
<bpepple> -- MARK -- Meeting End