Extras/SteeringCommittee/Meeting-20060914

2006 September 14 FESCo

Meeting Summaries are posted on the wiki at: http://www.fedoraproject.org/wiki/Extras/SteeringCommittee/Meetings

Attending

• thl
• jwb
• c4chris
• scop
• bpepple
• tibbs
• spot
• rdieter
• abadger1999
• dgilmore

Summary

Mass Rebuild

• Lot's of people didn't understand "bump Epoch-Version-Release" to mean EVR has to be incremented before rebuilding.
• perl packages should be rebuilt too. Only "*large* noarch content" such as game data is excluded.
• Packages not rebuilt will be removed from the download repository.

Extras Packages for Enterprise Linux aka EPEL aka Enterprise Extras

• Name will be Extras Packages for Enterprise Linux (by Fedora) aka EPEL
• RHEL will be used in the mock buildroots. Contributers can test on CentOS, there shouldn't be any difference between the two.
• Disttag will be .elX.
• Will EPEL start with RHEL4 or RHEL5?

How long will EPEL be supported

• RHEL is three years of feature adds and 4 years of security updates.
• EPEL packages will be supported for the full length of the targetted RHEL release.
• We would like to be able to hand off updating of older EPEL releases to another maintainer (in case the current maintainer no longer is running the older RHEL release.)
• Can a miantainer orphan a package or do they have to find a new maintainer first?
• EPEL is a rolling release (at least for now)
• Open question: Do we target the base RHEL release or the RHEL+updates releases?

Legacy in BuildRoots

• Will be added this weekend

Maintainer Repsonsibilities

• http://fedoraproject.org/wiki/Extras/Schedule/MaintainerResponsibilityPolicy

Packaging Committee Report

• Update to the pkgconfig guidelines: http://www.fedoraproject.org/wiki/Packaging Drafts/pkgconfig
• Directory ownership wording change: "Packages must not own files or directories already owned by other packages that they depend on. Exceptions to this rule are: perl...." with more detailed text explaining how and when perl packages are excluded from this
• New bylaw that a Package Committee member missing four meetings in a row without notifying the list is removed.

AWOL Policy

• To be discussed on the list.

Log

(10:00:29) thl has changed the topic to: FESCo Meeting in progress
(10:00:32) jima: ack!
(10:00:37) thl: Thursdays again already
(10:00:40) ***jwb is here
(10:00:44) thl: who's around?
(10:00:50) c4chris: thl, hi
(10:00:51) jima: blarney: i suppose we should quiet down for the duration of the meeting ;)
(10:00:52) scop [n=scop]  entered the room.
(10:00:56) blarney: jima: k
(10:00:56) ***bpepple is here.
(10:01:05) scop: yo
(10:01:06) thl: jwb, I might have to leave in half an hour or a bit later
(10:01:12) jwb: ok
(10:01:13) blarney: jima: message me off-channel
(10:01:14) ***jima is logging as rabble, in case anyone needs the log afterward.
(10:01:22) thl: jwb, could you finish the meeting in that case?
(10:01:26) jwb: thl, yep
(10:01:34) thl: jwb, tia
(10:01:44) tibbs: I'm here.
(10:01:47) ***spot is here
(10:01:52) spot: (for once)
(10:01:54) rdieter: here
(10:01:56) thl: so, welcome everybody!
(10:01:57) ***abadger1999 here
(10:02:10) thl has changed the topic to: FESCo Meeting in progress --  M{ae}ss-Rebuild
(10:02:17) thl: scop, any status update?
(10:02:40) scop: not here
(10:03:00) ***mmcgrath is here.
(10:03:05) scop: except that I'm flabbergasted how people didn't grok "bump the EVR"
(10:03:09) c4chris: still a ways to go...
(10:03:28) ***abadger1999 hit EVR with a stick, does that count?
(10:03:38) tibbs: scop: If it's not written down in bold type, most people will miss it.
(10:03:43) thl: scop, what about perl packages?
(10:03:48) thl: do they need a rebuild?
(10:03:48) tibbs: If it is written down in bold type, half of the people will miss it.
(10:03:56) thl: a lot of then were not rebuild yet afaics
(10:03:59) jwb: can't we add a script in CVS to check if release has been bumped or not?
(10:04:03) tibbs: Honestly, why not just rebuild everything and get it over with?
(10:04:06) sankarshan left the room (quit: "/me goes off to take a break").
(10:04:28) scop: noarch perl-* which install into the latest perl's versioned dir don't really benefit that much
(10:04:31) thl: tibbs, we need to find AWOL maintainers IMHO
(10:04:35) scop: (assuming they still build, of course)
(10:04:55) thl: well, should we enforce the rebuild for those packages?
(10:04:56) ***cweyl is lurking (in other meeting)
(10:05:00) tibbs: thl: Of course, that's one of the goals of having everything rebuilt.
(10:05:08) thl: scop, that might make stuff easier when we build stuff for RHEL5 later
(10:05:08) ***dgilmore is here
(10:05:15) tibbs: I wasn't implying that we (or an automated system) should do the rebuilds.
(10:05:28) thl: tibbs, ohh, sorry, then I misunderstood you comment
(10:05:47) thl: I really think we should poke all packagers
(10:06:02) thl: only stuff where it doesn't make any sense at all to rebuild should be left out
(10:06:03) scop: the original intention of leaving only *large* noarch content packages such as game date without a rebuild hasn't really happened
(10:06:04) dgilmore: scop: yeah  i dont get how people thought not bumping EVR  would work
(10:06:11) thl: e.g. large data packages
(10:06:20) c4chris: k, I'll poke again
(10:06:41) c4chris: all the ones that still have the needs.rebuild file
(10:06:41) thl: c4chris, please poke all packagers of noach packages, too
(10:06:54) thl: or does someone dislike that?
(10:07:02) bpepple: thl: Sounds good.
(10:07:18) c4chris: plus all those not rebuilt ?
(10:07:56) thl: c4chris, well, we probably have to manually look at those not rebuild yet where the needs.rebuild file was removed
(10:08:25) c4chris: thl, k, but that migth be a piece of work...
(10:08:27) thl: c4chris, maybe just mail them and tell them that only "*large* noarch content packages such as game data" don't need a rebuild
(10:08:38) c4chris: I'll try to produce some sort of list
(10:08:47) thl: c4chris, thx
(10:08:56) thl: k, anything else regarding the rebuild?
(10:08:58) abadger1999: Is the EVR problem big enough that we need a script to look into that?
(10:09:02) thl: what do we do after Sunday?
(10:09:18) thl: abadger1999, I don't think so -- maybe in the longer term
(10:09:18) dgilmore: thl: send emails and remove unbuilt packages
(10:09:28) scop: no
(10:09:31) scop: proceed as planned
(10:09:38) thl: dgilmore, that might rebult in a great mess
(10:09:44) scop: remove packages whose needs.rebuild is not taken care of
(10:09:46) tibbs: Unbuild packages should not be in FC6, though.
(10:09:52) jima: err, what about in the case of an AWOL maintainer?
(10:10:01) thl: scop, +1
(10:10:10) bpepple: scop: +1
(10:10:13) c4chris: scop, +1
(10:10:18) abadger1999: Better to break now than later +1
(10:10:20) spot: +1
(10:10:25) jwb: +1
(10:10:30) c4chris: jima, orphan process...
(10:10:36) thl: scop, will you take care of that after Sunday?
(10:10:39) tibbs: Is it reasonable to just not auto-branch them for FC-6?
(10:10:50) thl: tibbs, they need to be removed from the repo
(10:10:56) thl: otherwise they'll be in FE6
(10:11:01) scop: I can take care of that, yes
(10:11:10) jima: c4chris: whee. i'd better get moving on that.
(10:11:17) thl: scop, move then only aways -- just in case ;-)
(10:11:19) Rathann [n=rathann]  entered the room.
(10:11:27) scop: ha!
(10:11:46) scop: (I'll think about that ;))
(10:11:50) thl: scop, could you maybe send a small warning/heads up to the list before they get removed?
(10:11:56) thl: maybe today/tomorrow?
(10:12:03) ***Rathann sits quietly in the back row
(10:12:11) scop: will do
(10:12:19) thl: scop, tia
(10:12:21) tibbs: When you say "remove from the repo", are you deleting FC-5 and older branches as well?
(10:12:31) bpepple: tibbs: -1
(10:12:31) dgilmore: tibbs: no  just devel
(10:12:33) scop: poke me if I forget, I have a thousand things to do ATM
(10:12:35) thl: tibbs, no, only devel
(10:12:55) tibbs: OK.
(10:13:00) thl: we normally never remove stuff from released version
(10:13:04) thl: only devel
(10:13:11) thl: okay, anything else?
(10:13:19) thl: (regaing the mess rebuild)?
(10:13:38) bpepple: If a package is removed to not being rebuilt, how can it be re-added in the future.
(10:13:58) tibbs: We'll need a mechanism for requesting devel branches.
(10:13:58) dgilmore: bpepple: someone takes ove  and requests a branch
(10:14:01) nirik: what about dependencies?
(10:14:10) c4chris: bpepple, normal orphan process I guess
(10:14:13) liquidat [n=liquidat]  entered the room.
(10:14:20) bpepple: I've got a package that currently be built due to changed in the e-d-s api, that I haven't had a chance to fix.
(10:14:21) nirik: ie, something is needs.rebuild and has other items that already rebuilt using the older unrebuilt version?
(10:14:27) tibbs: We'll just have to sort out dependency problems as they happen.
(10:14:55) c4chris: tibbs, yes
(10:15:07) c4chris: they should be reported to f-e-l
(10:15:16) scop: we're not going to remove anything from CVS, just the devel package repo
(10:15:30) scop: s/just the/just from/
(10:15:35) dgregor [i=dennis]  entered the room.
(10:15:44) liquidat left the room ("Konversation terminated!").
(10:15:45) dgregor left the room (quit: Read error: 104 (Connection reset by peer)).
(10:15:47) scop: at least immediately, I think
(10:15:59) dgilmore: scop: that sounds right
(10:16:10) dgilmore: and add a dead.package file in cvs
(10:16:45) thl: dgilmore, we should re-visit this when we come closer to FC6/FE6
(10:16:58) thl: that's probably the best solution
(10:17:04) thl: that okay for everybody?
(10:17:08) dgilmore: sure
(10:17:12) c4chris: thl, agreed
(10:17:15) scop: yes
(10:17:16) abadger1999: ye
(10:17:21) thl has changed the topic to: FESCo Meeting in progress --  Enterprise Extras
(10:17:26) thl: k, let's move on
(10:17:32) thl: mmcgrath, z00dax, are you around?
(10:17:36) mmcgrath: pong
(10:17:45) thl: hi mmcgrath
(10:17:53) thl: just wanted to give you a heads-up
(10:18:07) ***mmcgrath is also troubleshooting our torrent system so I'm kind of here and not here.
(10:18:21) thl: well, after the last meeting I was wondering if we really need/want CentOS builders
(10:18:36) mmcgrath: What are we going to build them on?
(10:18:39) thl: wouldn't RHEL in the mock buildroots work better?
(10:18:43) jwb: wait
(10:18:48) jwb: can we agree on a name first?
(10:18:49) mmcgrath: does mock require yum?
(10:18:58) thl: people can test on centos if they want
(10:19:04) thl: there should be no difference
(10:19:14) thl has changed the topic to: FESCo Meeting in progress --  Enterprise Extras -- Name?
(10:19:18) tibbs: Isn't the name kind of the least important thing?
(10:19:33) jwb: tibbs, perhaps.  but it's something we _should_ be able to accomplish today
(10:19:34) mmcgrath: its also caused the most discussion.
(10:19:38) tibbs: It doesn't block any technical or organizational work.
(10:19:43) mmcgrath: I think the name depends on whether or not we consider this a branch or a fork.
(10:19:45) spot: Fedora RPMS: Enterprise Extras
(10:19:46) spot: FREE
(10:19:49) ***spot giggles
(10:19:49) tibbs: Fedora Extras.
(10:19:50) thl: FEEL
(10:19:58) jwb: Enterprise Extras
(10:20:11) thl: (Fedora Extras for Enterprise Linux)
(10:20:18) thl: or Fedora Extras
(10:20:25) thl: those get my vote
(10:20:31) mmcgrath: I personally like EE, nice generic and easy to understand.
(10:20:42) jwb: i'm fine with either EE or FE
(10:20:45) bpepple: +1 Fedora Extras.
(10:20:48) c4chris: FEEL is the most descriptive of what it actually is
(10:21:02) rdieter: FEEL++
(10:21:02) jwb: that's ok too i guess
(10:21:08) ***spot thinks having "Fedora" in it will confuse RHEL end-users... but doesn't care.
(10:21:14) mmcgrath: +1 to spot.
(10:21:17) tibbs: I'd really like to see Fedora in there.
(10:21:20) thl: spot, we can always rename it later
(10:21:21) jwb: i don't want RH* to be in the name
(10:21:22) mmcgrath: these packages will never make it on to a Fedora machine.
(10:21:24) bpepple: tibbs: +1
(10:21:25) thl: jwb, +1
(10:21:31) jwb: ok, FEEL+1
(10:21:43) ***dgilmore preferes EE
(10:21:55) spot: EE (from Fedora)
(10:22:16) abadger1999: spot: How about EE (by Fedora) ?
(10:22:22) spot: abadger1999: sure.
(10:22:23) bpepple: FEE?
(10:22:46) c4chris: abadger1999, ok too
(10:22:56) tibbs: I could get behind EE (by Fedora).
(10:23:07) jwb: or FEE
(10:23:20) scop: FREE or FEE? ;)
(10:23:20) spot: FEE: It'll cost you! ;)
(10:23:30) thl: I agree slightly with mschwendt when he said that EE sounds like "Software for Enterprise"
(10:23:37) ***mmcgrath doesn't understand why such pride exists over making sure Fedora gets into a package brand that will never be on a Fedora machine.
(10:23:58) tibbs: Fedora is the project I have volunteered so much of my time to.
(10:24:12) tibbs: My packages are Fedora packages.
(10:24:17) jwb: wait, stop
(10:24:35) jwb: lets settle on 3 options and then vote.  EE, FEEL, EE (by Fedora)
(10:24:43) c4chris: not necessarily pride, just honesty where they come from...
(10:24:58) bpepple: No FEE? :(
(10:25:14) abadger1999: bpepple:  it's neither fee as in beer, nor fee as in freedom. :-)
(10:25:15) spot: bpepple: no, it doesn't cost anything except bandwidth and time. ;)
(10:25:23) ***thl considers "Extras Packages for Enterrpise Linux"
(10:25:28) thl: EPEL
(10:25:34) dgilmore: i think it needs to be distro agnostic  as it will be used on RHEL and CentOS  and other rebuilds
(10:25:44) jwb: dgilmore, fair point
(10:25:56) jwb: thl's new suggestion sounds distro agnostic
(10:26:04) spot: Enterprise Extras Packages (EEP!)
(10:26:17) jwb: spot, stop polluting the name pool ;)
(10:26:24) spot: sorry. ;)
(10:26:34) thl: spot, same old problems; sound like software for enterprise usage
(10:26:35) bpepple: spot: Gotta have a little fun.
(10:26:35) abadger1999: spot: As long as the logo has a {{Template:Tip}}
(10:26:41) dgilmore: i could live with thl's suggestion
(10:26:47) thl: "Extras Packages for Enterrpise Linux (by Fedora)"
(10:27:05) jwb: does anyone _not_ like EPEL?
(10:27:14) spot: shouldn't there be a GNU in there somewhere? *ducks*
(10:27:27) c4chris: EPEL is fine with me
(10:27:30) ***spot likes EPEL fine
(10:27:31) ***thl hits spot with the stick
(10:27:36) rdieter: GNUEPEL
(10:27:42) dgilmore: EPEL +1
(10:27:42) rdieter: +1 (for any name, really, don't care, there's bigger fish to fry)
(10:27:49) abadger1999: EPEL fine
(10:27:52) jwb: tibbs, ?
(10:28:07) tibbs: I'm thinking.
(10:28:20) tibbs: I won't object to it, but I still would like Fedora in the name.
(10:28:21) ***jwb hums the jeopardy theme
(10:28:29) jima: that name seems to have a lot of EPEL...
(10:28:30) ***bpepple agrees with tibbs.
(10:28:31) ***jima hides
(10:28:52) mmcgrath: +1 epel
(10:29:11) jwb: epel +1
(10:29:14) ***spot is _starving_ ... brb, going to steal training food
(10:29:17) thl: epel +1
(10:29:19) jwb: i think that's a majority
(10:29:51) thl: jwb, agreed
(10:29:55) jwb: yay, we accomplished something today!
(10:29:56) jwb: :)
(10:29:59) mmcgrath: lol
(10:30:00) thl: we can still revisit it if poeple don#t like it
(10:30:05) jwb: right
(10:30:35) thl has changed the topic to: FESCo Meeting in progress --  Enterprise Extras -- build on centos or RHEL
(10:30:40) spot: eww. really bad chinese.
(10:30:41) thl: okay, back to this one
(10:30:49) mmcgrath: thl: honestly I'd be fine packaging both.
(10:30:58) dgilmore: thl: either is fine
(10:31:04) mmcgrath: I'm just unfamiliar with mock building RHEL
(10:31:06) bpepple: either seems ok.
(10:31:08) thl: I'd prefer RHEL
(10:31:09) spot: do we want a different dist tag for centos?
(10:31:14) jwb: thl, why?
(10:31:17) spot: we have established tags for RHEL
(10:31:17) thl: centos should always be compatible to RHEL
(10:31:21) rdieter: spot: imo, no.
(10:31:22) spot: but none for centos
(10:31:27) dgilmore: spot: nope  disttag of .el5
(10:31:28) thl: centos -> RHEL should work too
(10:31:36) mmcgrath: yeah, I'd say we could give these a dist tag similar to .el or .epel
(10:31:46) spot: ok, so if centos == RHEL, lets use RHEL. Red Hat will give us RHEL, i'm sure of it.
(10:31:58) dgilmore: i sya stick  with .el?  its already setup
(10:31:59) mmcgrath: Not to mention if the centos guys want branding, they're a build away from getting it :-)
(10:31:59) jwb: i can't test on RHEL though
(10:32:08) spot: jwb: but centos == RHEL
(10:32:18) spot: anything that differs is a centos bug
(10:32:19) thl: jwb, testing on centos should wrok, too
(10:32:19) spot: not a RHEL bug
(10:32:21) thl: or not?
(10:32:31) thl: (me has no access to rhel, only centos
(10:32:37) dgilmore: thl: same
(10:32:41) jwb: it doesn't bother me much
(10:32:50) mmcgrath: If packages don't cross-work then we should aid in correcting the OS.
(10:32:54) dgilmore: i think we can asuse for the most part binary compatability
(10:32:56) spot: mmcgrath: +1
(10:32:57) thl: mmcgrath, +1
(10:32:57) mmcgrath: in theory anyway :)
(10:33:05) rdieter: mmcgrath: +1
(10:33:10) spot: so, since RHEL is everyone's lowest common denominator...
(10:33:23) jwb: ok, that makes sense
(10:33:26) c4chris: spot, :)
(10:33:31) dgilmore: we use RHEL on builders
(10:33:41) dgilmore: easy to do
(10:33:46) thl: spot, can you help us getting a RHEL in a local yum-repo on the builders?
(10:33:46) jwb: dgilmore, we being EPEL or CentOS?
(10:33:53) thl: spot, with updates of course?
(10:33:55) dgilmore: jwb: EPEL
(10:33:56) spot: thl: sure.
(10:34:06) spot: thl: just tell me where you want it.
(10:34:08) jwb: k
(10:34:18) thl: spot: dgilmore and mmcgrath will know
(10:34:32) spot: i don't know either of them very well, but ok. ;)
(10:34:40) ***dgilmore slaps spot
(10:34:40) mmcgrath: hah
(10:34:53) scop: for how long are packagers expected to support RHEL releases?
(10:35:11) thl has changed the topic to: FESCo Meeting in progress --  Enterprise Extras -- how long are packagers expected to support RHEL releases
(10:35:18) thl: scop, good question
(10:35:23) tibbs: Probably the full seven years or whatever it is.
(10:35:27) scop: I'm interested in participating, but not for older EL/CentOS versions than what I use
(10:35:27) dgilmore: scop: as long as RH does   but we need a process to allow them to step down
(10:35:29) mmcgrath: actually we can get rhel copies from humphrey
(10:35:41) spot: the 7 years of RHEL isn't seven years of updates... its 3 years of feature adds, then the rest is just security
(10:35:54) thl: scop, I think we are talking about building for RHEL4 and later afaics
(10:36:01) scop: yes
(10:36:02) thl: or does anyone want to build for RHEL3?
(10:36:06) dgilmore: thl: i thought for RHEL5
(10:36:08) bpepple: thl: -1
(10:36:09) jwb: i think we should start with RHEL5
(10:36:09) thl: (jsut to be sure)
(10:36:15) c4chris: thl, nope
(10:36:23) thl: I think we should start with RHEL4
(10:36:24) scop: anyway, RHEL4 or not, that makes a fine example
(10:36:25) mmcgrath: I think http://www.redhat.com/security/updates/ should be authoritative
(10:36:33) tibbs: Packagers are going to have to think hard about whether they want to branch for EL at all.
(10:36:44) scop: I use 4 today, but will pretty quickly update everything to 5 when it's out
(10:36:49) thl: tibbs, +1
(10:36:54) tibbs: I imagine that some will only branch once someone shows up who wants to co-maintain on EL.
(10:37:02) jwb: tibbs, yes
(10:37:02) scop: after that I'm not personally too keen on working on 4
(10:37:11) mmcgrath: we also have xen boxes now so in theory when a developer can't test something we should be able to set an instance up for them, though that hasn't really been discussed as far as how possible it is.
(10:37:12) scop: does that preclude me from contributing?
(10:37:31) dgilmore: scop: i dont think so
(10:37:38) spot: scop: i say no, as long as you're willing to let someone else do any dists you don't want to
(10:37:51) scop: okay, good
(10:38:02) mmcgrath: Thats my question, what do we do when someone owns a package and doesn't want to create a branch for EL or for a specific version of it?
(10:38:03) scop: (of course I have nothing against others chiming in on earlier releases)
(10:38:04) c4chris: spot, sounds good to me
(10:38:13) tibbs: What architectures are we going to build on?
(10:38:25) dgilmore: mmcgrath: then it doesnt get done
(10:38:29) spot: tibbs: right now? x86, x86_64, ppc. ;)
(10:38:37) jwb: dgilmore, why?
(10:38:43) spot: unless someone has an s390 lying around.
(10:38:43) jwb: dgilmore, why couldn't someone else do it?
(10:38:44) ***thl has to leave now
(10:38:56) mmcgrath: later thl
(10:39:01) jwb: thl, ok i'll run
(10:39:01) dgilmore: jwb: if no one wants to branch for EL  then it doesnt branch
(10:39:03) thl: sorry guys; I'm sure you'll have a nice meeting without me
(10:39:06) c4chris: thl, later
(10:39:08) jwb: dgilmore, that wasn't the question
(10:39:15) dgilmore: jwb: someone else could if wanted
(10:39:24) jwb: dgilmore, ok that's what i was getting at
(10:39:49) mmcgrath: who's going to be topic mediator while thl's out?
(10:39:53) jwb: me
(10:40:27) jwb: so have we agreed that EPEL packages should be supported for the full RHEL release?
(10:40:36) bpepple: jwb: +1
(10:40:41) ***scop needs to go in 10 minutes
(10:40:42) rdieter: +1
(10:40:45) c4chris: jwb, +1
(10:40:48) tibbs: +1
(10:40:49) mmcgrath: +1
(10:40:51) jwb: +1
(10:40:53) mmcgrath: as much as possible.
(10:40:56) spot: ehh, why not. i wanna be owning lapack until i die.
(10:41:21) jwb: ok, so we've got full release support and building on RHEL
(10:41:23) spot: what about update releases of RHEL?
(10:41:33) spot: every quarter or so, RHEL respins itself with new packages
(10:41:46) c4chris: spot, we do rolling releases
(10:41:47) jwb: spot, do those get a new full 7 years?
(10:41:51) tibbs: There's no reason we have to sync to any release.
(10:41:53) mmcgrath: hmmmm.
(10:41:55) spot: jwb: no, we don't.
(10:42:05) mmcgrath: I'd be fine doing releases as well.
(10:42:06) spot: but these update releases might (shouldn't) break ABI/API
(10:42:23) mmcgrath: EPELu4
(10:42:33) rdieter: then deal with breakage on a case-by-case basis.
(10:42:33) jwb: spot, i think a rolling release strategy will cope with that
(10:42:37) spot: jwb: ok.
(10:42:41) c4chris: mmcgrath, EPEL-4u3 ?
(10:42:44) jwb: or does anyone disagree?
(10:42:55) rdieter: rolling++
(10:43:01) mmcgrath: EPEL4-u3,  Spot, they're really just snapshots right?
(10:43:11) c4chris: rolling++
(10:43:11) mmcgrath: so literally we'd be talking about making a copy of a point in time?
(10:43:23) dgilmore: rolling ++
(10:43:33) jwb: mmcgrath, snapshots for what?
(10:43:42) mmcgrath: for the specific releases of a redhat update.
(10:43:43) spot: mdomsch: yeah
(10:43:54) spot: err... mmcgrath: yeah
(10:43:56) jwb: mmcgrath, why are those needed?
(10:44:17) mmcgrath: I don't think they're needed really, but they would better comply with upstream.
(10:44:22) dgilmore: jwb: you could take a snapshot  and release a CD
(10:44:26) spot: RHEL might add a feature to a library in U2
(10:44:27) mmcgrath: We can revisit that later though.
(10:44:36) mmcgrath: Its trivial to add later.
(10:44:43) spot: a EPEL package decides it wants to use that
(10:44:47) jwb: i say we cross that bridge later then
(10:44:55) spot: ok.
(10:45:01) jwb: we don't do Extras CDs ATM either
(10:45:08) dgilmore: spot: then  they add BR's  etc  and build a new version
(10:45:15) ***mmcgrath is in an enterprise environment where the 'official releases' are a big deal.
(10:45:25) dgilmore: jwb: no  but we shoul when core can support them at install time
(10:45:39) jwb: dgilmore, i don't disagree.  but that's a different topic ;)
(10:45:44) dgilmore: yup
(10:45:52) jwb: dgilmore, and we can handle it then so both Extras and EPEL can cope
(10:46:00) jwb: or at least i would think
(10:46:09) dgilmore: jwb: yeah  i agree with you
(10:46:18) jwb: ok.  anything left for EPEL today?
(10:46:43) jwb has changed the topic to: FESCo Meeting in progress --  comps.xml
(10:46:50) mmcgrath: not right now, I have plenty of other questions to bring up later.
(10:46:52) jwb: c4chris, dgilmore: ?
(10:47:05) dgilmore: jwb: nothing to add right now
(10:47:07) c4chris: not much new
(10:47:07) jwb: mmcgrath, sure no problem.  just trying to get to some other stuff today :)
(10:47:07) mmcgrath: brb, lunch
(10:47:14) ***mmcgrath ^z
(10:47:19) jwb: ok.  anyone else on comps?
(10:47:33) dgilmore: jwb: nope
(10:47:43) jwb has changed the topic to: FESCo Meeting in progress --  Activate legacy in buildroots
(10:47:53) jwb: dgilmore, you're up again :)
(10:48:12) dgilmore: jwb: it will be done  by endf of weekend
(10:48:17) dgilmore: end of
(10:48:28) dgilmore: ive  cleared up some things  and will get it done
(10:48:35) jwb: ok good.  is anyone working on the maintainers responsibilities part?
(10:49:04) jwb: i take that as a no
(10:49:08) tibbs: Well, there was the draft, but I haven't received any comments.
(10:49:18) jwb: tibbs, where was that again?
(10:49:29) abadger1999: tibbs: Hey -- I posted :-)
(10:49:38) tibbs: Let me see if I can find it agai.
(10:49:43) abadger1999: (Too the list, not the wiki)
(10:49:57) tibbs: http://fedoraproject.org/wiki/Extras/Schedule/MaintainerResponsibilityPolicy
(10:50:01) jwb: dgilmore, and you will document in the wiki the FE branches in maintenance mode use Legacy packages?
(10:50:22) jwb: tibbs, can you add a link to that in the status section?  or do you think it needs to be it's own topic?
(10:50:35) dgilmore: jwb: it will be done
(10:50:44) tibbs: I thought thl had done so when he moved it under Extras/Schedule.
(10:50:44) jwb: dgilmore, excellent
(10:51:06) jwb: tibbs, i don't see it
(10:51:12) ***scop needs to go now, seeya
(10:51:15) tibbs: Honestly I think it should be its own topic.
(10:51:23) jwb: tibbs, i agree.  want to add it?
(10:51:30) tibbs: Yes, I'm in there now.
(10:51:38) jwb: great
(10:51:45) jwb: ok, anything else on this?
(10:52:01) jwb has changed the topic to: FESCo Meeting in progress --  PC report
(10:52:08) jwb: anything from the PC this week?
(10:52:09) c4chris: scop, later
(10:52:17) jwb: bye scop
(10:52:19) tibbs: We did make some progress.
(10:52:52) tibbs: Unfortunately it really needs to be written up in a presentable format.
(10:53:09) tibbs: We have to get these meetings separated in time a bit.
(10:53:09) spot: ok, so here are the items that we handled
(10:53:24) jwb: tibbs, understand
(10:53:40) spot: We approved an update to the pkgconfig guideline.
(10:53:42) spot: PackagingDrafts/pkgconfig
(10:54:11) spot: We also approved a change in wording around Directory ownership:
(10:54:20) spot: "Packages must not own files or directories already owned by other packages that they depend on. Exceptions to this rule are: perl...."
(10:54:48) spot: with more detailed text explaining how and when perl packages are excluded from this
(10:55:32) spot: And: We passed a rule that if a PC member misses four meetings in a row (without notifying the list), they're removed.
(10:55:33) jwb: ok
(10:55:41) jwb: oh
(10:55:43) jwb: ok
(10:55:45) spot: They get a warning at three.
(10:55:59) jwb: has that been a problem?
(10:56:05) spot: unfortunately, yes.
(10:56:19) jwb: well i'm glad to see FESCo hasn't hit that yet :)
(10:56:25) jwb: spot, ok.  anything else?
(10:56:32) spot: nope. thats a lot for us. :)
(10:56:41) xris [n=xris]  entered the room.
(10:56:44) jwb: ok.  thanks
(10:56:46) jima: might not be a bad idea to vote in that idea for FESCo.
(10:57:01) jwb: jima, perhaps.  but not this week
(10:57:08) jwb: ok, we've got 1 minute
(10:57:10) bpepple: jima: It doesn't seem to be a problem so far.
(10:57:16) jima: bpepple: so far.
(10:57:29) jwb: is there anything people want to discuss that needs to be done this week?
(10:57:40) tibbs: AWOL Policy.
(10:57:49) tibbs: But we can do this on-list.
(10:58:01) jwb: tibbs, yes let's do it on-list this week
(10:58:09) jwb: thanks for emailing that out
(10:58:28) jwb: ok.  i have to leave now, but you guys can keep going
(10:58:30) jwb: who wants the topic baton?
(10:58:44) pygi left the room (quit: Read error: 110 (Connection timed out)).
(10:58:46) jwb: or i'll ajourn in 30
(10:58:55) bpepple: ajourn.
(10:59:02) tibbs: Sorry, I'm buried in wikicode.
(10:59:12) tibbs: I have nothing else to add.
(10:59:26) abadger1999: nothing from me either.
(10:59:27) jwb: ok.  it seems like a good stopping point anyway
(10:59:35) ***jwb closes the meeting
(10:59:50) jwb: -- MARK --