From Fedora Project Wiki
< Extras | SteeringCommittee
(10:00:27 AM) thl has changed the topic to: FESCo meeting -- Meeting rules at [WWW] http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Init process
(10:00:28 AM) jima: i don't mind doing dependency-hunting for other people
(10:00:31 AM) thl: FESCo meeting ping -- abadger1999, awjb, bpepple, c4chris, dgilmore, jeremy, jwb, rdieter, spot, scop, thl, tibbs, warren
(10:00:34 AM) thl: Hi everybody; who's around?
(10:00:34 AM) ***jima hushes up
(10:00:35 AM) kwizart: i also cannot run squid cache... but i'm working on...
(10:00:37 AM) ***jeremy is around
(10:00:37 AM) rdieter: yo
(10:01:04 AM) ***abadger1999 is here
(10:01:09 AM) c4chris|w: hey gang
(10:01:15 AM) c4chris|w left the room (quit: Remote closed the connection).
(10:01:17 AM) ***nirik is sitting in the rabble bleachers with his coffee.
(10:01:42 AM) ***dgilmore is here
(10:01:51 AM) ***thl counts tibbs, rdieter, jeremy, c4chris, dgilmore, /me
(10:01:57 AM) thl: k, then let's start
(10:01:59 AM) c4chris|w [n=chris] entered the room.
(10:02:03 AM) ***|DrJef| is sitting inhis MST3K styled seating ready to make snarky comments
(10:02:04 AM) thl has changed the topic to: FESCO-Meeting -- EPEL -- mmcgrath, dgilmore -- status update
(10:02:07 AM) ***spot is here
(10:02:15 AM) thl: mmcgrath, dgilmore ?
(10:02:28 AM) scop [n=scop] entered the room.
(10:02:41 AM) dgilmore: thl: i pinged notting again on getting the sync setup
(10:02:55 AM) dgilmore: still waiting from word back
(10:03:06 AM) c4chris|w: any news about me setting up an ia64 builder ?
(10:03:12 AM) dgilmore: I need to get bugzilla setup
(10:03:16 AM) dgilmore: c4chris|w: not yet
(10:03:28 AM) nirik: are all the epel packages now in the epel owners list?
(10:04:01 AM) ***rdieter thinks all his are there now.
(10:04:12 AM) dgilmore: nirik: i need to audit that
(10:04:13 AM) c4chris|w: nirik: dunno. Mine are...
(10:04:19 AM) thl: nirik, mine are, too
(10:04:29 AM) tibbs: At least denyhosts isn't.
(10:04:53 AM) ***jima vaguely notices his aren't...oops.
(10:04:54 AM) thl: but it seems many people still are not aware that epel is now open for them, too -- we need to annouce that propaply
(10:04:57 AM) nirik: clamav is also not, and I saw a build of it.
(10:05:09 AM) c4chris|w: I'll try to make a version of my package status script to check EPEL stuff
(10:05:12 AM) thl: and we really need rules -> latest and greates, more stable, mix?
(10:05:24 AM) thl: c4chris, that would be helpful
(10:05:31 AM) jima: thl: yes, i didn't know it was ready for public consumption.
(10:05:40 AM) nirik: also would be good to have a broken deps report for it on push like extras...
(10:05:43 AM) tibbs: clamav in epel is also old and vulnerable to at least one attack.
(10:05:48 AM) dgilmore: thl: hopefully we can get things moving faster now the Christmas/ New Year is behind us
(10:05:49 AM) thl: jima, it's for sponsors and people with more than 5 packages
(10:06:03 AM) jima: thl: i fall into the latter category, yes.
(10:06:13 AM) thl: dgilmore, EPEL seems to be still lacking a driver that takes care of all the shit
(10:06:31 AM) thl: dgilmore, you and mmcgrath seem to be busy with a lot of other stuff already
(10:06:35 AM) thl: I'm, too
(10:06:35 AM) dgilmore: thl: I guess that will be me
(10:07:04 AM) dgilmore: thl: i will make time to be more proactive with EPEL
(10:07:19 AM) thl: dgilmore, thx
(10:07:55 AM) thl: hmm, seems I can skip the next two topics then, as we slightly discussed them already
(10:08:03 AM) thl has changed the topic to: FESCO-Meeting -- Opening Core -- rdieter, jeremy, warren -- status update
(10:08:04 AM) rdieter: do we have any public-consumable epel repo yet?
(10:08:32 AM) thl: rdieter, jeremy, warren ?
(10:08:33 AM) dgilmore: rdieter: not yet. unless you grab the unsigned rpms off of the build server
(10:09:01 AM) rdieter: (sorry missed last board meeting)
(10:09:11 AM) jeremy: the final i has been dotted and so the process of pushing people towards a) ensuring they have accounts in the account system and b) starting to get their packages reviewed should be starting this afternoon
(10:09:25 AM) rdieter: rock.
(10:09:39 AM) jeremy: (this is as of ~ 15 minutes ago :-)
(10:09:52 AM) jeremy: had hoped to get that prior to the holidays, but just couldn't nail it down :/
(10:09:54 AM) thl: nice
(10:10:15 AM) f13: wooo~
(10:10:15 AM) thl: so what do we need to do to make everything running smoothly?
(10:10:27 AM) f13: help get people Fedora accounts
(10:10:33 AM) jeremy: build system bits I now need to follow up about and make sure things roll roll roll along, but that's a whole separate discussion and conversation
(10:10:37 AM) f13: figuring out how to sponsor a ton of RH folks will be.. interesting.
(10:11:04 AM) thl: f13, that's exactly one of the crucial points...
(10:11:33 AM) thl: any idea how to solve that?
(10:11:47 AM) dgilmore: f13: i will sponsor some of them
(10:11:59 AM) thl: should we distribute the red hat people over all the sponsors?
(10:12:03 AM) f13: thl: I honestly don't know :/
(10:12:05 AM) jeremy: thl: generally speaking, the purpose of the sponsorship is accountability, yes?
(10:12:10 AM) thl: they should be easy to handle mostly
(10:12:12 AM) f13: is there anything that actually tracks sponsorship?
(10:12:14 AM) tibbs: I've sponsored many Red Hat folks in the past.
(10:12:18 AM) rdieter: I nominate jeremy to sponsor everybody. (;
(10:12:18 AM) thl: jeremy, yes
(10:12:22 AM) nirik: so theres 1500ish packages and how many rh people?
(10:12:22 AM) ***jeremy sponsors people often
(10:12:23 AM) |DrJef|: thl, we are also need to have some sort of targets for how fast on average core packages need to make it through review to get everything done before fc7 release
(10:12:29 AM) nim-nim left the room (quit: Remote closed the connection).
(10:12:34 AM) thl: nirik, good question
(10:12:36 AM) tibbs: f13: there's a way to extract a list from the account system.
(10:12:36 AM) f13: or would be be able to just blanket say "RH management will be responsbile and sponsor all these new RH folks"
(10:12:41 AM) f13: ah.
(10:12:42 AM) tibbs: Perhaps someone knows the magic URL.
(10:13:01 AM) f13: |DrJef|: mass reviews could be somethign that happens at the FudCon
(10:13:08 AM) f13: mass reviews+fixes that is
(10:13:14 AM) f13: because many many packages will need fixes
(10:13:15 AM) jeremy: my thinking the past with RHT people is that I sponsor them; if they do something wrong, I hammer their management
(10:13:15 AM) thl: |DrJef|, see f13#s answer
(10:13:16 AM) abadger1999: f13: Are we concerned that any RH packagers will take shortcuts when reviewing fellow RHer's packages?
(10:13:23 AM) abadger1999: Just to get the bits into the repository?
(10:13:27 AM) nirik: https://admin.fedoraproject.org/accounts/dump-group.cgi?group=cvsextras&role_type=sponsor&format=html
(10:13:27 AM) jeremy: (this is what I've been doing in the past)
(10:13:40 AM) thl: jeremy, I know, and that's okay for me, but some people complained about that to me in private :-/
(10:13:54 AM) jeremy: abadger1999: same issue as the discussion about checklists -- the only way to avoid things like that is peer review
(10:14:08 AM) tibbs: There's also https://admin.fedoraproject.org/accounts/dump-group.cgi?group=cvsextras&format=html but it doesn't show who sponsored who.
(10:14:37 AM) nirik: will we be using bugzilla/the regular review process?
(10:14:53 AM) abadger1999: jeremy: I agree. With that in mind I'd be against a blanket sponsorship of RHers.
(10:14:54 AM) jeremy: thl: longer term, I want the answer to be that new RHT engineers are sponsored by their "peer" as part of the standard indoctrination/orientation stuff. but that won't work for this case
(10:14:55 AM) jeremy: nirik: yes
(10:14:56 AM) f13: abadger1999: if that happens, engineers get the bat as do managers.
(10:14:58 AM) c4chris|w: nirik: yes
(10:14:58 AM) thl: jeremy, any rough number about how many rh packagers we are talking?
(10:15:17 AM) nirik: cool.
(10:15:24 AM) abadger1999: Or alternately: Maybe we split sponsored status now so there's a level that can not do reviews.
(10:15:39 AM) jeremy: thl: off the top of my head estimate -- maybe 100-ish?
(10:15:45 AM) daMaestro left the room (quit: "Leaving").
(10:15:49 AM) thl: jeremy, k, thx
(10:15:54 AM) jeremy: abadger1999: anyone can do reviews, though; even people who aren't sponsored. they just can't _approve_ the reviews
(10:16:00 AM) jeremy: thl: I can get a more accurate number later this afternoon
(10:16:01 AM) c4chris|w: so maybe we each sponsor 10 ?
(10:16:07 AM) thl: well, we probably need to bring this discussion to the list once
(10:16:23 AM) thl: c4chris|w, well, I'd say each sponsor 5?
(10:16:31 AM) thl: I think we should have 20 active sponsors iirc
(10:16:34 AM) nirik: you also need fedorabugs to be able to change bugzilla buts to review things...
(10:16:39 AM) c4chris|w: thl: that'd be fine
(10:16:41 AM) abadger1999: jeremy: Ah. Sorry. forgot that all RH'ers will have the equivalent of fedorabugs.
(10:16:56 AM) jeremy: nirik: everyone @redhat.com has the equivalent of fedorabugs
(10:17:04 AM) nirik: yeah... makes sense.
(10:17:30 AM) thl: k; anyway, we need to discuss this once in the public on the list
(10:17:43 AM) thl: any volunteers that want to kick of the discussion?
(10:17:50 AM) jeremy: thl: I'll start a thread once bill has sent the other mail
(10:17:58 AM) thl: jeremy, k, thx
(10:18:11 AM) thl: I'll move on
(10:18:12 AM) thl has changed the topic to: FESCO-Meeting -- Opening Core -- rdieter, jeremy, warren -- FESCo future? Deadlock? [WWW] https://www.redhat.com/archives/fedora-advisory-board/2007-January/msg00003.html)
(10:18:29 AM) thl: I feel a bit lost
(10:18:50 AM) thl: I'd like to get a bit more input from the board what they think how FESCo should be run in the future
(10:19:11 AM) dgilmore: thl: i agree we need a little guideance as to whats expected
(10:19:54 AM) thl: jeremy, rdieter, should we have once a special meeting with both FESCO and the Board?
(10:19:57 AM) nim-nim [n=nim-nim] entered the room.
(10:19:58 AM) ***jeremy is 100% agreed that "what does governance look like in our new world" isn't entirely clear (also, I hate the word governance ;)
(10:20:19 AM) thl: or at least certain members of both groups?
(10:20:23 AM) rdieter: I agree a meeting of the minds is in order.
(10:20:42 AM) f13: another con-call
(10:20:49 AM) f13: because IRC is teh suq for that kind of thing
(10:20:56 AM) thl: one meeting to brainstorm, then someone writes a proposal, and then we discuss this further on the list and/or another meeting
(10:20:58 AM) c4chris|w: how about during next FESCo meeting?
(10:21:03 AM) jeremy: I think the big thing is boiling things down to what the real questions are: 1) what responsibilities exist? 2) who (what person/group of people) is accountable to the specific responsibilities and 3) how do we do these without ending up with a bazillion committees
(10:21:33 AM) thl: f13, you don't know my spoken english yet ;-)
(10:21:37 AM) rdieter: Personally, I have no problem simply with FESCo (re)defining itself (I can't think of anyone better suited to the task).
(10:21:51 AM) c4chris|w: teh suq?
(10:22:03 AM) jeremy: rdieter: *nod* I tend to agree that some amount of fesco input here is not only desired, but really required
(10:22:09 AM) jeremy: much like how the packaging committee came about
(10:22:20 AM) |DrJef|: c4chris|w, translation.... irc really sucks
(10:22:36 AM) thl: jeremy, the packaging committee is part of the problem IMHO
(10:22:39 AM) c4chris|w: |DrJef|: thx
(10:22:56 AM) jeremy: thl: I don't disagree
(10:22:57 AM) thl: we had some problems in the past where it was unclear if a task need to be handled by FESCo or by the PC
(10:22:59 AM) f13: c4chris|w: sorry, "geek" speak for "it is not very good"
(10:23:49 AM) thl: well, I have a bit of free time tomorrow and on sunday; i can try to think about hte whole issue a bit and post something to the list
(10:23:52 AM) rdieter: ok, maybe we can see about getting more Board folks to sit on the next FESCo meeting, does that sound do-able?
(10:24:03 AM) jeremy: rdieter: sounds reasonable
(10:24:08 AM) thl: but I'm not yet sure if I really find the time for it
(10:24:19 AM) jeremy: thl: I'll also try to write some stuff up (I've thought about it a bit, just haven't had much time to write things up)
(10:24:21 AM) thl: rdieter, yeah, might be a good idea
(10:24:34 AM) thl: jeremy, the usual problem afaics ;-)
(10:24:36 AM) abadger1999: PC isn't really beneath FESCo in the hierarchy right now. We'll have to decide if that should change in the new world.
(10:24:36 AM) ***jeremy will make sure someone else drives to work tomorrow and work on it during the commute ;)
(10:24:58 AM) thl: abadger1999, I think that should change; it should be beneath FESCo in the hierarchy
(10:25:04 AM) ***rdieter nods.
(10:25:13 AM) f13: jeremy: won't be me, I have a 9am concall
(10:25:16 AM) abadger1999: thl: It would clarify things quite a bit.
(10:25:37 AM) rdieter: How about, we say it is, until proven otherwise? (:
(10:25:43 AM) thl: abadger1999, it whould aslo help if we clearly document what group is responsable for what tasks
(10:26:03 AM) abadger1999: And without a Core/Extras split there's no need for it to speak to two different audiences.
(10:26:13 AM) thl: abadger1999, yes, exactly
(10:26:31 AM) ***c4chris|w nods
(10:26:33 AM) thl: anything else? otherwise I'll move on
(10:26:49 AM) thl has changed the topic to: FESCO-Meeting -- Encourage co-maintainership -- c4chris
(10:26:49 AM) abadger1999: If we hav ediscussion at next FESCo meeting
(10:27:00 AM) ***thl waits a bit more
(10:27:01 AM) abadger1999: We should plan on not having much time for other business.
(10:27:23 AM) rdieter: makes sense.
(10:27:27 AM) thl: abadger1999, yes, probably; but that should be no problem if there are no other pressing issues
(10:27:28 AM) c4chris|w: yup
(10:27:43 AM) thl: and normally there aren't any
(10:27:53 AM) thl: anyway, now for co-maintainership
(10:27:58 AM) thl: I have to say sorry here
(10:28:15 AM) thl: I wanted to write something up during christmas/new year, but I did not find the time...
(10:28:16 AM) ***c4chris|w sheepishly admits to haven't done much for co-maintainership...
(10:28:27 AM) tibbs: Well, it's been happening on its own anyway.
(10:28:42 AM) thl: parts of, yes
(10:28:50 AM) tibbs: Perhaps not enough, but at least a few people have been signing up as co-maintainers.
(10:28:59 AM) thl: c4chris, do you plan to work on it in the next days? otherwise I'll try
(10:29:11 AM) c4chris|w: not in the next days, no
(10:29:20 AM) thl: k, then I'll try
(10:29:26 AM) thl: anyway, a bit of input would be nice
(10:29:33 AM) thl: what would you guys like to see?
(10:29:42 AM) thl: a primary-maintainer and co-maintainers?
(10:29:55 AM) thl: or "all maintainers are equal"?
(10:29:56 AM) c4chris|w: I think yes
(10:30:10 AM) c4chris|w: no, there should be one reference
(10:30:32 AM) rdieter: I tend to think we prefer to have a primary,go-to person.
(10:30:34 AM) Belegdol: all maintainers are equal, mou some are more
(10:30:36 AM) ***nirik likes the idea of primary and co-maintainers, and they can divide work however works for them, but the primary is the responsible party
(10:30:39 AM) Belegdol: s/mou/some
(10:30:54 AM) thl: nirik, +1
(10:31:08 AM) c4chris|w: nirik: +1
(10:31:12 AM) thl: and how many co-maintainers do we want to encorage? 2?
(10:31:13 AM) rdieter: nirk: +1
(10:31:18 AM) tibbs: Yes, let's leave structure up to the maintainers.
(10:31:19 AM) thl: encourage
(10:31:33 AM) thl: still wrong :-/
(10:31:49 AM) rdieter: I think it should enough to encourage folks to have at least 1 co-maintainer.
(10:32:02 AM) ***thl votes for two
(10:32:19 AM) tibbs: It wouldn't be bad to identify packages which need more, like network-facing servers or things we use for infrastructure.
(10:32:33 AM) nirik: perhaps on packages that are critical and/or have a bad security history/are kmods the security or a kernel group to be automatic co-maintainers.
(10:32:52 AM) rdieter: 2 is probably a bit overkill for small/simple packages, maybe too little for big/complex ones.
(10:32:53 AM) abadger1999: thl: FYI In the pkgdb I'm designating one person as primary maintainer, but he can grant all of his rights to comaintainers as well.
(10:32:53 AM) thl: tibbs, yeah, might be a good idea (but we should not limit the number of co-maintainers in any case)
(10:32:57 AM) c4chris|w: packages with a security-hole history should get more co-maintainers
(10:33:19 AM) f13: there also may need to be maintainership on a per collection basis
(10:33:30 AM) nirik: abadger1999: or just a few?
(10:33:32 AM) f13: one person may want to maintain the Fedora side of things, while another maintains EPEL
(10:33:41 AM) abadger1999: nirik: Correct.
(10:33:45 AM) thl: f13, that's for sure afaics
(10:33:55 AM) f13: the new buildsystem code tracks package "ownership" on a per collection basis, as it tracks packages themselves for a per collection basis
(10:33:59 AM) abadger1999: f13: Yes. Ownership is assigned to package-in-collection
(10:34:06 AM) f13: while not really all that possible with owners.list, in a DB this becomes easy.
(10:34:07 AM) rdieter: f13: Hmm... probably, could we get away with just considering them *all* co-maintainers?
(10:34:32 AM) f13: rdieter: possibly, but per collection ownership is pretty important
(10:34:38 AM) Belegdol left the room (quit: "Leaving").
(10:34:46 AM) cwickert left the room (quit: Read error: 60 (Operation timed out)).
(10:34:47 AM) f13: especially when this information is being used for whack-a-maintainer mode.
(10:34:51 AM) rdieter: f13: esp to (re)assign primary maintainer..
(10:35:03 AM) f13: we have a bot internally that we can do 'whoowns foo' or 'whoowns foo <collection>' to talk to the owner
(10:35:13 AM) rdieter: nice. (:
(10:35:23 AM) thl: I think that the EPEL maintainer normally should be at least a co-maintainer for fedora (and the other way around, too)
(10:35:45 AM) f13: forcing it the other way around isn't very nice
(10:35:55 AM) cwickert [n=chris] entered the room.
(10:36:00 AM) thl: f13, that why I said "should" ;-)
(10:36:05 AM) f13: as a Fedora maintianer, I may have violent objections to being on the hook for anything RHEL related.
(10:36:07 AM) thl: we really should not force it
(10:36:10 AM) f13: k
(10:36:33 AM) thl: k; anything else regarding co-maintainership?
(10:36:46 AM) rdieter: f13: being comaintainer mostly only Cc's you for bugs, which, imo, would be a good thing. What else would you consider bad?
(10:36:54 AM) nirik: if a owner can adjust how they delegate at any time that should be good enough... then the RHEL maintainer can remove a fedora co-maintainer that doesn't want anything to do with it.
(10:37:38 AM) f13: rdieter: maintaines are on the hook for security updates. If one maintainer isn't available, the default next is the co-maintainer. I may not want to know anything about how the package operates in RHEL, I only care about Fedora.
(10:37:39 AM) rdieter: I guess per-collection owners will give us the best of both worlds. I can live with that.
(10:37:40 AM) thl: abadger1999, can people in the package database CC to bugs without being a real co-maintainer ?
(10:37:56 AM) nirik: with the new system, there would be things like cvs acls, being able to do builds?
(10:38:00 AM) |DrJef|: f13, i for one can't commit to co-maintaining EPEL branches at the moment
(10:38:00 AM) thl: s/to bugs/to packages/
(10:38:08 AM) abadger1999: thl: yes. I've got an ACL table per package-in-collection.
(10:38:25 AM) thl: abadger1999, k, great :)
(10:38:32 AM) abadger1999: And there are separate ACLs for bugzillamail, commitmail, commit, change packagedb, etc.
(10:38:41 AM) [splinux] [n=splinux] entered the room.
(10:38:46 AM) c4chris|w: I think packages mentioned in fedora-security/audit/f* should have a couple co-maintainers
(10:39:23 AM) thl: getting late
(10:39:27 AM) thl: I'll move on
(10:39:28 AM) thl has changed the topic to: FESCO-Meeting -- MISC -- firefox updates in stable often breaks quite some packages (galeon, liferia, ...); do we need some kind of task force that steps up to rebuild the stuff?
(10:39:39 AM) f13: abadger1999: seperate ACLs or seperate ACL systems? One would hope not seperate systems. :/
(10:39:46 AM) |DrJef|: thl, i guess i should comment since i brought it up to the lists this time
(10:39:56 AM) thl: |DrJef|, shoot
(10:39:58 AM) tibbs: nirik and I have been doing a lot of rebuilds to fix broken deps.
(10:40:01 AM) |DrJef|: thl, 2 comments
(10:40:05 AM) f13: thl: in the new world, this would be caught by depchecking. and yeah, a FF SIG might be appropriate.
(10:40:07 AM) abadger1999: f13: One table, separate ACLs.
(10:40:09 AM) rdieter: imo, simple, firefox maintainer should notify maintainers list, and use updates-testing (more).
(10:40:27 AM) |DrJef|: thl, 1) we need a versioned build requires for sure against firefox... so we get broken dep messages from the scripts
(10:40:30 AM) f13: rdieter: or even better, the update tool would block the update from going out so long as there are broken deps
(10:40:37 AM) rdieter: f13: that too. (:
(10:40:41 AM) |DrJef|: thl, 2) this also affects at least one package in core
(10:40:42 AM) f13: much of this problem goes away with xulrunner though I hear.
(10:40:52 AM) |DrJef|: thl, so an Extras sig isnt good enough
(10:40:54 AM) nirik: but for security thats all bad. You want the update asap.
(10:41:09 AM) f13: nirik: yes, but we don't push security updates if there are broken deps internally either
(10:41:17 AM) f13: because you wouldn't be able to install the security update
(10:41:30 AM) f13: we just have more packages to worry about now, a fully staged update
(10:41:31 AM) |DrJef|: rdieter, not happening.....how does he know what the full list of dependant packages are?
(10:41:34 AM) tibbs: But we can't coordinate between extras and core now.
(10:41:40 AM) thl: |DrJef|, +1 for some kind of meanism that makes sure the we get broken dep messages from the scripts
(10:41:47 AM) tibbs: We can't rebuild extras package until the update has been released.
(10:41:51 AM) nirik: f13: agreed.
(10:42:14 AM) rdieter: |DrJef|: reqoquery is your friend. Besides, doesn't matter, notice should be sent to fedora-maintainers.
(10:42:15 AM) kushal [n=kd] entered the room.
(10:42:20 AM) |DrJef|: thl, im also concerned about this..beyond firefox.... do we have other versioned library trees out there?
(10:42:22 AM) ***nirik is looking forward to the review of the 'esc' package... it's gonna be a doozy.
(10:42:22 AM) thl: f13, the stuff lmacken is working on should fix that iirc
(10:42:27 AM) f13: rdieter: sometimes thats not possible if its for embargo issues
(10:42:30 AM) |DrJef|: rdieter, you mis understand
(10:42:38 AM) f13: thl: more to the point, having all the packages in one repo would fix this.
(10:42:43 AM) |DrJef|: rdieter, without a versioned dep on firefox.... repoquery isnt all that useful
(10:42:50 AM) thl: f13, hehe, yes :)
(10:42:52 AM) |DrJef|: rdieter, because the automatically library deps.......
(10:43:00 AM) |DrJef|: rdieter, still show clean on a packaging level
(10:43:06 AM) f13: right, repoquery doesn't help, as versions may stay the same, but the app may still break
(10:43:19 AM) thl: f13, ?
(10:43:24 AM) f13: Chris clearly needs minions and communication
(10:43:31 AM) rdieter: |DrJef|: Ah, the versoined dep's are definitely needed (I assumed dependancies were already taken care of completely)
(10:43:37 AM) f13: thl: ff libraries may not change version, but may become incompatible.
(10:43:40 AM) |DrJef|: rdieter, no thats part of the problem
(10:43:54 AM) nirik: |DrJef|: esc is the only package that has that issue you have found right?
(10:44:01 AM) |DrJef|: rdieter, the libraries down in the versioned firefox tree.... rpm catches them as deps...but not the pathname
(10:44:08 AM) thl: f13, but normally that happens only on a package update with a newer version
(10:44:10 AM) |DrJef|: nirik, no esc was actually bogus
(10:44:17 AM) thl: I'd like to fix this normally
(10:44:17 AM) |DrJef|: nirik, python-gstreamers-extras is a real one
(10:44:25 AM) f13: anywho, I think we all agree
(10:44:28 AM) alleycat left the room (quit: Remote closed the connection).
(10:44:28 AM) ***rdieter is glad not to have to mess with firefox ('n friends) packaging. Sounds not fun at all.
(10:44:35 AM) f13: lets get some volunteers to help Chris with coordination when possible.
(10:44:38 AM) |DrJef|: nirik, esc is fucked up...
(10:44:47 AM) |DrJef|: nirik, it totally disabled dep checking in its spec file
(10:44:55 AM) thl: f13, and a list of packages depending on firefox somewhere in the wiki
(10:44:57 AM) |DrJef|: nirik, that package is going to get beatdown on review
(10:44:58 AM) f13: but as stated, may not always be possible due to embargo.
(10:45:02 AM) Foolish left the room (quit: Remote closed the connection).
(10:45:02 AM) nirik: |DrJef|: yes, that would never pass review...
(10:45:32 AM) thl: I also think someone should just rebuild the "depnding on firefox" pacakges if a new firefox gets pushed
(10:45:44 AM) thl: we should not have to wait for each and eevry individual maintainer
(10:45:45 AM) |DrJef|: thl, pipedreaming..... if rpm knew that libraries in certain locations needed to have their location tracked....
(10:46:05 AM) |DrJef|: thl, but for the packages that don't depend on firefox explicitly?
(10:46:16 AM) thl: |DrJef|, hehe, rpm is complicated enough already IMHO, but yes, you hae a point...
(10:46:27 AM) |DrJef|: thl, we'll need a script which watches for all the library provides from firefox as well
(10:46:51 AM) rdieter: |DrJef|: firefox packaging seems a bit screwed to me then, if things really are that location dependant, then firefox shouldn't be moving around, and just use (something like) /usr/lib/firefox
(10:46:55 AM) thl: |DrJef|, well, the maintainers should normally know if their package depends on firefox
(10:47:00 AM) thl: let them make a list in the wiki
(10:47:09 AM) |DrJef|: rdieter, it is screwed... the maintainer agrees.... firefox 3 should be much better
(10:47:15 AM) thl: and let somebody rebuild all the packages from that list when a new firefox gets pushed
(10:47:20 AM) |DrJef|: rdieter, i think upstream will be splitting out the libraries from the interface
(10:47:30 AM) |DrJef|: rdieter, xulrunner...specifically
(10:47:38 AM) rdieter: yay, light at the end of a (long) tunnel.
(10:47:44 AM) |DrJef|: thl, minus the core packages :->
(10:47:52 AM) thl: |DrJef|, sure
(10:48:06 AM) thl: |DrJef|, in the merged world it IMHO should be chris job to do that
(10:48:10 AM) |DrJef|: rdieter, the larger question is.... are their other versioned trees out there....that break less frequently..but in the same way?
(10:48:12 AM) thl: or one of his co-maintainers
(10:48:32 AM) rdieter: |DrJef|: no doubt, probably. (:
(10:48:47 AM) |DrJef|: thl, you'll be hard-pressed to get chris to do it... but a co-maintainer sure
(10:49:23 AM) thl: |DrJef|, well, we could give chris a script "run this after it got pushed"
(10:49:31 AM) |DrJef|: rdieter, i'm racking my tequila soaked brain thinking of a way to find those potential problem packages that need explicit versioned requires
(10:49:48 AM) |DrJef|: thl, also firefox breaks the crap out of readahead still i think
(10:50:04 AM) nirik: |DrJef|: python-gstreamers-extras is where? core? extras? I don't see it off hand.
(10:50:07 AM) |DrJef|: thl, but that's another issue
(10:50:10 AM) |DrJef|: nirik, core
(10:50:10 AM) thl: |DrJef|, I'd like to move on; could you keep an eye on this stuff and driver it forwards so the situations at least gets a bit better for everyone?
(10:50:20 AM) |DrJef|: nirik, listen in extras requires it for example
(10:50:32 AM) |DrJef|: thl, drive it how?
(10:51:15 AM) |DrJef|: thl, i've done pretty what I can do.. identifying the problem... i can try to make a script to identify all affected packages using repoquery
(10:51:23 AM) thl: well, what you suggested earlier for example; " some kind of meanism that makes sure the we get broken dep messages from the scripts if a new firefox got out"
(10:51:31 AM) |DrJef|: thl, and file bugs to get firefox versioned deps in
(10:51:53 AM) |DrJef|: thl, this will need to be enshrined in review policy
(10:52:11 AM) thl: |DrJef|, let's chat another time about it and move on now
(10:52:13 AM) |DrJef|: thl, i'll work on spinning up some scripts
(10:52:16 AM) |DrJef|: thl, good enough
(10:52:22 AM) thl has changed the topic to: FESCO-Meeting -- MISC -- "what do we want like to see in the approved-message in a review bug" -- [WWW] https://www.redhat.com/archives/fedora-maintainers/2006-December/msg00214.html
(10:52:27 AM) thl: It's quite late already
(10:52:35 AM) thl: and another flamewar fest :-/
(10:52:45 AM) thl: do we want to skip it for today?
(10:52:47 AM) c4chris|w: for sponsors: I don't care
(10:52:55 AM) c4chris|w: for newcomers: some details...
(10:53:03 AM) thl: c4chris, +1
(10:53:27 AM) tibbs: All I have to add is that I wonder how we should determine who to make a sponsor if it's acceptable to have a review with no details.
(10:53:30 AM) rdieter: common sense rears it's ugly head, c4chris, +1
(10:53:38 AM) abadger1999: I think newcomers should be aware that they stand a much better chance of being sponsored/evevated to sponsor status if they provide full details in their review.
(10:53:53 AM) thl: abadger1999, +1
(10:54:03 AM) c4chris|w: precisely
c4chris_ c4chris|w
(10:54:08 AM) abadger1999: c4chris: +1
(10:54:26 AM) nirik: reviews with no details could also be pushed into the 're-review' queue before others (although I have no idea how you would identify them off hand).
(10:54:29 AM) thl: what about license and checksum?
(10:54:36 AM) |DrJef|: thl, and reviewers should indudge submitters if they ask for more detailed comments
(10:54:48 AM) thl: |DrJef|, sounds fine to me, too
(10:55:03 AM) abadger1999: thl: Still subject to spot check :-(
thimm thl
(10:55:15 AM) |DrJef|: thl, to build personal trust between submitter and reviewer
(10:55:21 AM) scop: I don't think there are many "perfect" packages being submitted, almost everything has some issues that a reviewer should catch or ask an explanation for
(10:55:25 AM) abadger1999: thl: Although fedora.us used to have some automated checking of checksum, right?
(10:55:37 AM) tibbs: We have to periodically check things anyway. Look at the XaraLX issue for a good reason why.
(10:55:43 AM) |DrJef|: scop, would it shock you that i delibrately add errors to my specs... to test my reviweer?
(10:55:56 AM) ***rdieter snickers
(10:56:00 AM) c4chris|w: |DrJef|: evei :-)
(10:56:08 AM) c4chris|w: evil too
(10:56:11 AM) thl: abadger1999, the bugzilla messages were gpg signed and had to contain the md5sum of the pacakge
(10:56:15 AM) f13: so long as dumping the guidelines into the bug isn't made manditory, I don't care (:
(10:56:21 AM) nirik: tibbs: yeah, would that we have enough people stepping up to do reviews that we could have more than one person review each package. ;(
(10:56:50 AM) Tjikkun__ is now known as Tjikkun
(10:56:54 AM) thl: anyway, let's leave it with that now and move on
(10:56:55 AM) c4chris|w: we'll soon have 20 more reviewers, right ;-)
(10:56:57 AM) |DrJef|: f13, i still think my side-bar issue has merit.... finding better ways to inform contributors when policy guidance changes in a granular way
(10:56:58 AM) nirik: |DrJef|: have any you introduced been missed by your reviewer? :)
(10:57:02 AM) thl has changed the topic to: FESCo meeting -- Packaging Committee Report -- spot, abadger1999, rdieter, tibbs, scop
(10:57:07 AM) |DrJef|: nirik, no...
(10:57:19 AM) thl: I've seen no report on the list...
(10:57:20 AM) |DrJef|: nirik, not the ones i delibrately added :->
(10:57:22 AM) tibbs: We just discussed tweaking the PHP guidelines.
(10:57:24 AM) nirik: cool. tats encouraging. ;)
(10:57:35 AM) f13: |DrJef|: absolutely, having a ever updated checklist and a 'whats changed recently' type thing is good.
(10:57:50 AM) tibbs: We're adding some hints, no rules are actually changing.
(10:58:01 AM) thl: tibbs, k
(10:58:07 AM) thl has changed the topic to: FESCo meeting -- Sponsorship nominations -- MichałBentkowski vel. ecik
(10:58:11 AM) tibbs: However, once we work with the RH PHP maintainer some additional changes are coming.
(10:58:30 AM) thl: MichałBentkowski +1 from me
(10:58:30 AM) tibbs: I gave my +1 on the list for ecik.
(10:58:36 AM) c4chris|w: +1
(10:58:38 AM) abadger1999: +1
(10:58:58 AM) thl: bpepple gave a +1 on the list, too
(10:59:03 AM) thl: warren also
(10:59:27 AM) thl: that makes six +1
(10:59:28 AM) |DrJef|: nirik, sorry its gnome-python2-extras thats got the firefox problem
(10:59:32 AM) thl: nobody yelling
(10:59:38 AM) thl: approved
(10:59:45 AM) ***scop doesn't know ecik's work well enough to have anything but a mild positive feeling
(10:59:57 AM) Rathann: he's been doing a lot of reviews
(11:00:04 AM) ***rdieter has a warm, fuzzy feeling.
(11:00:43 AM) thl has changed the topic to: FESCo meeting -- Maintainer Responsibility Policy -- bpepple
(11:00:48 AM) thl: ohh, not around
(11:00:51 AM) thl: skipping
(11:00:55 AM) thl has changed the topic to: FESCo meeting -- Free discussion around Fedora Extras
(11:01:06 AM) nirik: I have one item I would like to bring up...
(11:01:09 AM) nirik: https://www.redhat.com/archives/fedora-maintainers/2007-January/msg00006.html
(11:01:30 AM) thl: ohh, yes, forget about that
(11:01:53 AM) nirik: I doubt maintainers will close all those bugs...
(11:02:02 AM) rdieter: +1 blanket close, and be done with it.
(11:02:05 AM) thl: I'm all for "This release is no longer supported, can you retest with fc6 and re-file if it is still an issue'
(11:02:11 AM) c4chris|w: I think someone did a mass close for previous releases ?
(11:02:22 AM) abadger1999: I like davej's messages.
(11:02:23 AM) nirik: yeah, but I can't recall who it was...
c4chris_ c4chris|w
(11:02:30 AM) abadger1999: c4chris: I think it was nodoid
(11:02:31 AM) nirik: yeah, davej's messages rock...
(11:02:41 AM) ***thl like the davej messages, too
(11:03:19 AM) abadger1999: Shall we do that? Copy his message, set to NEEDINFO, and close after two weeks if there's no movement?
(11:03:21 AM) nirik: shouldn't it be someone in redhat? or I guess we could be mean and use the 'nobody@fedoraproject.org' address?
(11:03:29 AM) thl: abadger1999, +1
(11:03:40 AM) scop: while taking care of my < FC5 bugs, I noticed some old ones I'd already forgotten that had "devel" as the version
(11:03:41 AM) c4chris|w: abadger1999: +1
(11:03:56 AM) nirik: abadger1999: sounds great to me... who will do it tho? :)
(11:04:31 AM) abadger1999: Wonder if davej automates that...
(11:04:32 AM) ***nirik listens to the crickets chirp.
(11:04:37 AM) scop: so it might not hurt to go through bugs reported against devel while at it, if someone has the energy
(11:04:54 AM) thl: scop, isn't that the maintainers job?
(11:05:10 AM) f13: notting did it for Fedora/RHL bugs in the past.
(11:05:13 AM) f13: ping him
(11:05:19 AM) c4chris|w: I can give that a shot
(11:05:36 AM) mether [n=ask] entered the room.
(11:05:39 AM) scop: thl, I personally think that it's the maintainer's job to take care of all his bugs, but since people are discussing that some other folks would do it...
(11:05:41 AM) nirik: scop: how would we know on devel? when it was filed?
(11:05:43 AM) abadger1999: Are there really only 76 bugs for FE3/4?
(11:05:49 AM) thl: scop, maybe we should tell maintainers to look at all the bugs below <certain number>, as most of them are probably not up2fate anymore
c4chris_ c4chris|w
(11:05:57 AM) xris [n=xris] entered the room.
(11:06:10 AM) scop: nirik, devel and bug id < N (N about 170000 or something)
(11:06:13 AM) nirik: abadger1999: thats what my search had. Might have done it wrong... thats open/needinfo, etc)
(11:06:18 AM) thl: scop, +1
(11:06:31 AM) ***scop giggles at thl's up2fate ;)
(11:06:31 AM) abadger1999: c4chris|w: I can help too.
(11:06:43 AM) c4chris|w: abadger1999: k
(11:06:46 AM) thl: scop, :)
(11:06:52 AM) thl: had not noticed that typo
(11:07:18 AM) daMaestro [n=jon] entered the room.
(11:07:27 AM) nirik: also on the subject of bugzilla, is there any way to speed it up? ie, archive off all the pre fedora redhat bugs or something? it's got to be getting bloated... oh well, thats off topic I'm sure.
(11:07:27 AM) c4chris|w: thl: you're very good at funny typos :-)
(11:07:42 AM) thl: so, who will take care of those old bugs?
(11:07:46 AM) nirik: c4chris|w: If I can assist too let me know.
(11:07:51 AM) abadger1999: Do we want to give maintainers 'til Monday or just do it ourselves? With only 76 bugs, I can see two people doing it manually this time around.
(11:08:00 AM) Karlik [n=Karlik] entered the room.
(11:08:00 AM) ***c4chris|w and abadger and nirik
(11:08:08 AM) abadger1999: three :-)
(11:08:39 AM) thl: abadger1999, can we do this "change several bugs and once" feature for needinfo/closing?
(11:08:43 AM) ***thl is confused
(11:08:54 AM) c4chris|w: abadger1999: why not shoot an email to f-e-l an dwait 'til next week
(11:09:06 AM) thl: c4chris|w, fedora-maintainers imho
(11:09:17 AM) c4chris|w: thl: oh, right
c4chris_ c4chris|w
(11:09:24 AM) abadger1999: c4chris: +1
(11:09:48 AM) thl: c4chris|w, +1
(11:09:56 AM) c4chris|w: I'll post something after the meeting
(11:10:03 AM) thl: c4chris|w, k, thx
(11:10:07 AM) thl: anything else?
(11:10:12 AM) nirik: then we wait a week, move them all to needinfo and then another week and close?
(11:10:23 AM) c4chris|w: nirik: yes
(11:10:27 AM) nirik: right. ok.
(11:10:33 AM) thl: nirik, +1
(11:10:54 AM) ***nirik has no other items to bring up
(11:11:07 AM) ***thl will end the meeting in 60
(11:11:44 AM) ***thl will end the meeting in 30
(11:12:06 AM) ***thl will end the meeting in 10
(11:12:16 AM) thl: -- MARK -- Meeting End