FESCo Meeting 2007-07-26

--- bpepple has changed the topic to: FESCo meeting -- Meeting rules at http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Init process
* tibbs here
poelcatbpepple: i might not be able to be here for the feature portion
bpepplethat's ok.  Is the Feature Dashboard current?
poelcatbpepple: yes :) please go through the "ready for approval" section
bpepplepoelcat: cool.  should be no problem then.
poelcati've got to check in at the Fedora booth... if they have enough coverage, i'll try to find access
there is no (free) wireless coverage on the exhibit floor at OSCON
bpepple:(
* poelcat biab
bpeppleFESCo meeting ping -- bpepple, caillon, c4chris, dgilmore, dwmw2, f13, jeremy, jwb, notting, spot, nirik, tibbs, warren
* spot is here
dwmw2fish
bpeppleHi everyone, who's around?
* notting is here
caillon...
* jeremy is here
rdieterhere.
* warren here
rdieteroops, goes to sit in rabble seats now... :)
* rdieter shares popcorn with knurd
knurd busy fixing deps in EPEL
knurd can need popcorn for that
bpeppleok, let's get started...
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- elect FESCo Chair - all
bpeppleIs anyone interested in being in chair besides jwb and me?
* spot drops a pin
dwmw2 takes that as a 'no' :)
caillon doesn't want to be sat on, so no
bpeppleok, then let's do a quick vote.  either 'jwb' or 'bpepple'.
* c4chris here
tibbsWhen you ask for volunteers to stand in front of the firing squad, you don't usually get many.
c4chrisbpepple: +1
caillon(abstain)
tibbsbpepple: +1
spotbpepple (nothing against jwb, but I'm inclined to let the current dictator continue)
f13Im here.
dwmw2yeah, likewise. bpepple +1
* wwoods lurks
jeremybpepple: +1
f13bpepple: +1 just for he's been doing a good job, and jwb_gone didn't really sound like he /wanted/ the job, just that he'd do th ejob.
notting bpepple +1
c4chrisf13: yup, that was my impression too
bpeppleok, that looks like seven '+1' for me, so it looks like I'll be the chair.
ok, next item....
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- Feature Proposal Approval - http://fedoraproject.org/wiki/Features/Dashboard - notting, poelcat
bpepplepoelcat said he probably won't be here for this part, so I'll go ahead and lead this.
First item that needs our approval is: http://fedoraproject.org/wiki/Releases/FeatureBuildId
tibbsDidn't we approve many of the "ready for approval" items?
nottingtibbs: these are new ones
afaik
poelcatbpepple: i'm back
bpepplenotting: correct, these are new ones.
caillonso before we do this, i'd like to know what we're trying to vote on.  are we going to tell people to stop doing work, for example if we vote no?
bpepplepoelcat: great, I'll let you lead on this then. ;)
caillonwhat does voting mean
tibbsI'm somewhat confused about FeatureBuildId.
dwmw2yeah, I'm a little confused how this correlates with the use of F8Target for RFE bugs. Normally, the rule has sort of been "get it working in time and it gets in".
poelcathttp://fedoraproject.org/wiki/Releases/Features/NodokaTheme
no response from art-list
f13caillon: AFAIk a vote no is usually attached with a reason, as in "your plan sucks" or "please don't do this feature" or "not this release"
warrenlooks like it would require a mass rebuild
tibbsA mass rebuild wouldn't be a terrible thing.
dwmw2are we going to start with the Nodoka Theme?
f13caillon: depending on the reason is what we would tell the developers.  It could be 'revise your plan a bit and try again' or it could be 'try for next release' or it could be 'that's not legal dude, no way'
dwmw2or was that it? :)
tibbsBut when Roland asked some questions in fedora-packaging about this, he presented it as a fait accompli.
f13warren: I'm not afraid of a mass rebuild, so long as it's for something like this.
poelcatcaillon: it is all here: http://fedoraproject.org/wiki/Features/Policy#head-03985d7eeb69f08288d3116d96f3d4512ac3556b
* warren notices that we opted not to go with gcc-4.2 yet
f13poelcat: if we haven't gotten a response from art-list, perhaps we need to ping them again and tellthem this is their last chance for input on the matter.
dwmw2warren: we probably don't have time for that kind of fun if we want F8 out so soon.
bpepplef13: yeah, we should contact Marin since I believe she leads the art team.
jeremyI can take the lead on doing so
* poelcat will contact marin
jeremyshe's back from guadec now so in the office
caillonand if we say no, and someone does it anyway what recourse do we have?
tibbsI thought we were discussing FeatureBuildID.
warrendwmw2, yeah, breaks C++ ABI for the first time since 3.3.x or so
f13and on the bus most days, so able to get attention from.
bpepplespeak of the devil. there's mizmo.
caillon(i also think it would be useful to spend some time with welcome to the new guys, and explaining some things)
f13caillon: untag builds, block contributor access
* mizmo holds out pitchfork
c4chrisI see no real problem with FeatureBuildID, so +1
nottingmizmo: fesco would like to confirm that nodoka is the art team's plan for f8
mizmo: along with echo
mizmo: see http://fedoraproject.org/wiki/Releases/Features/NodokaTheme
mizmonotting: while efforts are being made towards echo, i dont think there are many on the art team who realistically think it will be ready for f8, especially with the shortened timeframe
notting: we are pretty stretched timewise just for the basics like wallpaper, gdm, etc
f13mizmo: are the owners of this feature on the art team?
mizmonotting: nodoka is definitely being planned to be released in time for f8, but of course we have no authority to make any decision about it being default
nottingyou don't?
mizmof13: which? nodoka or echo?
f13I'm a bit confused as to why owners would be trying to drive something that the team doesn't thin kwill happen.
mizmonotting: no, it's the same story as f7.
f13mizmo: well, both really.  Doesn't nodoka depend on echo?
mizmof13: what do you mean by owner?
f13Name: DanielGeiger, MartinSourada
mizmo: every proposed feature has an owner, the person responsible for making it happena nd reporting status and all that
mizmof13: while the visual style of nodoka was designed with complementing echo in mind, i don't see any reason why nodoka couldn't be used with mist
f13ok.
mizmof13: martni sourada is, i'm not 100% sure about daniel geiger (At least in terms of being on the art team accordnig to the account system)
f13: echo is being driven by luya rather than martin or daniel
f13mizmo: if these guys go forward with nodoka does that conflict with any Fedora Art plans?
(putting aside which icon set it'd use for the moment)
mizmof13: nope, it doesn't
f13: i think they've been doing great work
f13ok, that's good for me.
mizmof13: were there any specific concerns about it? just the echo <=> nodoka relationship?
nottingmizmo: mainly just making sure that nodoka was part of the art team's plan
mizmoah okay cool
bpepplemizmo: we just wanted to make sure that the art team was ok with if becoming the default theme.
f13mizmo: we don't want another tear with the Art team.
all together now!
mizmoah okay. well we haven't actually discussed it being default since we've pretty much unhappily resigned to the fact we have no authority over defaults. if there was any possibility of us having that, i think folks on the team would probably want to test it out and give it a final look over before it was officially sanctioned as the default (does that make sense?)
wwoods"no authority over defaults"? who does hold that authority, if not the art team?
mizmowwoods: desktop team
caillonmizmo: i think the relevant engineering teams (desktop in this case) would be happy to do the engineering work to make it happen if there is a clear plan that is agreed upon
* spot thinks it makes sense to let the Fedora art team have that authority
caillonbasically, the art team just needs to say "do this" authoritatively
and we really have no objections there
* drago01 agrees with spot
bpepple+1 to having the art team having the authority on themese.
jeremy+1
* abadger1999 agrees with spot... but has everyone looked over the history of the issue?
c4chris+1 (but dunno the history)
dwmw2I haven't, which is why I kept quiet. Are you going to enlighten us? :)
mizmowwoods: part of the reason the f8 theming process was delayed is i kept trying to get some kind of answer about this, and i could get no guarantees from the relevant parties that we'd be given any authority. so when w estarted the f8 theming process, i wrote a note to the team letting them know there were no guarantees their hard work would end up getintg used, but i promised to try to at least get them shipped as alternatives as best
i could
spotabadger1999: i suspect it is legacy of a pre-merge Fedora.
* caillon doesn't think that there is any authoritiy to be given
mizmoi just felt really bad going ahead announcing the f8 theming process, telling people to work on something that might not get used you know
spotwell, afaik, FESCo has the authority to say that.
mizmocaillon: well you can think what you want. it's not true.
f13spot: actually I think a lot of the 'history' came from last release.
spot+1 for the art team being able to determine the art.
f13f7 or fc6
caillonmizmo: i don't see how anyone on desktop team can even claim that as we don't have anyone working on art.
f13but regardless of history, I'm all for Art team having authority over themeing and such, so long as they work /with/ the Desktop teams and such, but I suspect that's a no-brainer.
mizmocaillon: whatever.
wwoodsI know I'm a bit naive about the process but isn't the default theme basically set by a string in a package somewhere? So long as the theme meets some agreed-upon requirements it should be trivial to change, right?
* f13 suspects that there might have been something with Diana involved.
dwmw2on the whole it would seem to make sense for the art team to be responsible for the art, if we're going to have an art team at all :)
* mizmo is having some blood pressure issues, if you dont need me any more then ill be off
rieldwmw2: +1
caillonagreed
warrenf13, there was a bit of drama involved back then when the desktop team asserted authority and everything blew up.  We're trying to start with a clean slate now.
* spot agrees with dwmw2 (how often does that happen? ;)
c4chrisdwmw2: exactly
* f13 looks for a second moon
dwmw2spot: more often than not. It's just less fun so we remember is less vividly :)
f13mizmo: i think that's all we need.  Thanks!
drago01mizmo: not being the default theme does not mean not used at all... what about seperate packages?
bpepplemizmo: thanks for your time.
dwmw2spot: we'll get to kmod (spit) later, and I'll agree with f13 too :)
mizmodrago01: there are technical issues which much of the theming that makes it difficult to easily swap themes out
later
mclasenray is working on cleaned up, split artwork packaging
f13so back to the feature ?
+1 feature.
mclasenI hope he'll get back to that soon
dwmw2so, since we don't have 'approval' from the art team
nottingiirc, nodoka was already approved pending art team ok
dwmw2perhaps we should give them another week to look over it with the expectation that we _will_ listen to them?
f13notting: ah ok.
warrenWe need a clear and binding statement on who is accountable and has authority over art.
caillondwmw2: sounds fine to me
warrenAnd it isn't clear that we have the authority to make it happen.
f13dwmw2: based on what I heard from mizmo, I don't think there is /disapproval/ which is th eimportant thing.
bpeppledwmw2: that sounds fine with me.
nottingwarren: fedora project board :)
dwmw2they seem not to have understood that we were actually asking for that.
abadger1999FESCo should probably let the Nodoka authors know that we'll go with whatever the art team decides.
warrennotting, there is a lot of unstated history here
* f13 read that as unsalted history
nottingwarren: and... not going into that here.
abadger1999So they can continue working.
mclasenas far as the desktop team is concerned I have stated pretty clearly on the art-list that we are not going to be involved with that for f8
* poelcat wonders if we are ready for the next feature?
spotbpepple: perhaps we should vote on letting the art team be the authority for Fedora art, reporting to FESCo?
c4chrisspot: agreed
poelcatspot: where is "the art team" defined?
dwmw2I'm a little wary of voting on such a thing not understanding the history, when there clearly _is_ history. But I can just abstain :0
spothttp://fedoraproject.org/wiki/Artwork
warrenIf the desktop team has said they wont be involved, then it seems that there is no opposition to art team having authority.
As long as we're clear on that point, then I see no problem here.
bpepplespot: Yeah, that sounds like a good idea to me. let's do a quick vote.
warrenbpepple, what exactly are we voting on?
* spot is having lag issues, i'm getting irc in chunks here
bpeppleProposal: Art team has authority for Fedora art, and report to FESCo,
spot+1
dwmw2ok then, let's just state that the art team are responsible for artwork and ask them for clear approval (or otherwise) of the plan to use Nodoka for F8
tibbs+1
bpepple+1
c4chris+1
warren+1
dwmw2+1
f13+1
warrenNodoka is a separate issue than authority
notting+1 until we get mapplethorped
jeremy+1
bpeppleok, that's nine '+1', so that's approved.
caillonlet me guess, the engineering guys do engineering.  and the docs team does docs, too.  /me not sure why this is a vote, but +1
dwmw2warren: yes, it is. But having given them authority, we should ask them to _use_ it :)
bpepplecaillon: agreed, but the art team seems to think they don't have authority over the art.
warrencaillon, you know clear why they feel that way.
jeremycaillon: something like that :)
dwmw2well, now we can tell them they do.
and move on to BuildID, perhaps?
bpeppleok, we should probably move on if we plan to cover more of the features. ;)
poelcat: ?
poelcathttp://fedoraproject.org/wiki/Releases/FeaturePresto -- need to confirmation from infrastructure team that Fedora can handle requirements; posted question to fedora-infastructure-list--no clear answer of "yes, we can do it".
bpepplepoelcat: let's poke the infrastructure team some more, and get a definitive answer.
c4chrisso let's skip this one?
jeremypoelcat: I think the infrastructure side isn't "we can't do them".  just a matter of getting the buildsys code done in time
bpepplec4chris: +1
poelcatbpepple: okay will leave for next time
spotmmcgrath: is infrastructure ok with Presto?
poelcatnext is #
next is http://fedoraproject.org/wiki/Releases/FeatureBuildId
#
c4chris+1 (looks like it's mostly in anyway)
f13also we need some more clear ideas of what kind of space this is going ot take up, and whether or not our current netapps can handle it.
wwoodsIIRC patches for buildid are already in binutils
f13but that also depends on when garbage collection in koji lands which would give us a lot more space
jeremythere are a few packaging things which need to be verified out with BuildID, but everything is in progress
f13I'm +1 for buildID, including a mass rebuild at some point.
(to be determined)
warrenHow long will a mass rebuild take?
f13depends
if rel-eng gets all BOFH about it and just scripts the crap out of it, a week or so. If we want maintainers to do it themselves, 3 months.
c4chriswe can ask mdomsch
dwmw2+1 to buildid
f13(a bit overblown of course)
bpepple+1 to buildid, here also.
tibbsbuildid sounds OK to me, but I wonder how this is going to interact with versioned -debuginfo packages as we have them currently.
warren+1 buildid
notting+1 buildid
spot+1
dwmw2and to doing it automatically, then taking anyone who screams "you touched my package" out back and quietly shooting them in the head
f13c4chris: mdomsch's numbers are just pure 'how long does it take to compile'.  Not at all counting the cvs spec editing time, the fixing of borken build times, the coordination with maintainers times, koji's load times.
c4chrisf13: ah, right, there's that too :)
tibbsdebuginfo packages aren't parallel-installable, so what does this extra infrastructure actually help with?
wwoodsBuildID provides some interesting value for Apport - it means we don't have to do any sort of depsolving to determine which debuginfo packages a given corefile/binary needs
well. apport and debugging in general
bpeppleI count seven '+1', so far.  anyone else want to weigh in before moving on?
poelcatNext feature for discussion is: http://fedoraproject.org/wiki/Releases/FeaturePolicyKit
tibbsIs anyone here who is actually driving this feature?
* mmcgrath here
tibbsSorry, I was asking about buildid.
jeremytibbs: the buildid stuff?  roland is.  and I've talked with him some about it
wwoodsI've talked with him about it as well
poelcattibbs: that is the definition of a feature "owner"
mmcgrathspot: do you still need me?
tibbsBecause I've not been able to get an answer to the question I asked.
poelcateach feature has a designated one
hmm
spotmmcgrath: i don't think so.
jeremytibbs: what it gains is a much easier way of mapping from binary -> debuginfo that will be less processor intensive and that should eventually allow us to do duplicate dwarf info pruning (which will mean smaller debuginfo) and some other things like that
mmcgrathspot: k
f13still no contengency plan
* mmcgrath goes back to lunch :)
tibbsGah, it's pointless to try and discuss these things on IRC.
cailloni'm not sure what we're talking about now anyway
jeremytibbs: difficult at least, yes :)
dwmw2caillon: http://fedoraproject.org/wiki/Releases/FeatureBuildId
f13poelcat: this one still seems to be missing the end user impact section as well.
dwmw2caillon: oh, I lied. sorry.
f13dwmw2: we moved on from that, we're now no Policy Kit
notting+1 on policykit from me
caillonoh
bpepplesome of the policykit items are still in package review aren't they?
wwoods(jeremy: can you ping roland to get some answers for tibbs and fleshing out the missing sections? is he in "boston"?)
nottingbpepple: yes
f13looking at the feature page I'm finding it hard to figure out how this is differnet than say consolehelper
tibbsI like the idea of policykit, but is it close enough to being done?
f13(and yes, I'm being a bit 'difficult' here)
nottingf13: it's intended to replace it.
dwmw2f13: see also http://fedoraproject.org/wiki/Releases/FeatureRemovePAMConsole :)
mclasenpolicykit made a big jump forward yesterday, with david releasing a new version
jeremytibbs: well, as hal is about to be depending on it, we don't have a whole lot of choice :-/
tibbsErm, then why are we even talking about it?
bpepple+1 on policykit.
dwmw2+1
caillontibbs: i wondered about some of these things earlier  (not sure why we need to vote on things like this)
+1 though
dwmw2are we getting hal 0.5.10 for f8 then? There were other things bluetooth-related which were waiting on that
f13notting: might be worth noting that in the feature page (:
mclasenyes, we are
c4chris+1 policykit
f13+1
dwmw2caillon: yeah, this sounds like one of the cases where they should just get on with it
nottingf13: it's picemeal. it's not going to replace it all at first drop. but that's the long-term intention
bpeppledwmw2: agreed.
nottingcaillon: ideally, this has a 'how to fix your app to use PK' somewhere, then we publicize it as a feature to point people to how to take advantage of it
wwoodsI'd love a "PolicyKit and ConsoleKit: what they do and how they differ" doc somewhere.
caillonnotting: which is a different issue.  but i agree that would be nice.
tibbshttp://people.freedesktop.org/~david/polkit-spec.html is the closest thing, I guess.
caillonnot sure if david'z already done that
* spot idly wonders how PolicyKit hooks into SELinux
caillonwith teeth
spotno, thats how my cat hooks into SELinux.
nottingwwoods: not a doc, but CK is for tracking who's logged in where. PK is a framework for what user X is allowed to do
tibbsIt's mentioned briefly at the above URL, I guess.
nottingPK can use CK as one of its criteria
wwoodsnotting: gotcha, thanks
mclasenCK is a utmp replacement of sorts
caillondavidz had some presentations somewhere about it, wwoods
i'll see about getting him to blog about it or something
esp the newest releases
* bpepple looks at the time, and sees we only have 10 minutes left. :(
f13send more votes
jeremy+1
warren+1
nottingok, that's 8 +1
dwmw2I thought we had enough already
poelcatNext feature for discussion is: http://fedoraproject.org/wiki/Releases/FeatureApport
bpepple+1
dwmw2why do we need to vote on this? It's not even doing to be installed or enabled by default
jeremywwoods: not wanting to enable it by default?
wwoodswee! I've actually got a FESCo question for this - do I go forward with the tg-based crashdb that's separate from bugzilla, or use the rh bugzilla as the backend?
spotopen source serverside code from ubuntu? *boggles*
wwoodsdisabled by default, yes
I'd suggest that once it's more mature we have it installed and enabled by default for test releases
tibbsYou should probably avoid bugzilla.
nottingwwoods: then why not target F9 with it?
jeremywwoods: aha, okay
LetoTowasnt bugizlla going away?
caillonwwoods: do not use bugzilla
dwmw2+1 to Apport. Please don't make us vote on _every_ new package which isn't installed or enabled by default :)
wwoodsokay - I'll need to discuss the crashdb with the infrastructure team
since we need somewhere to send reports
caillonwwoods: the mozilla guys have gotten this done right.  put it in a web database like crash-stats.mozilla.org
nottingif this isn't going to be a default-for-f8 thing, I say we punt it as a 'Feature'
abadger1999wwoods: You can catch me after the meeting.
f13dwmw2: this isn't about voting for new packages, it's about touting it as a Feature that can be advertised, blogged about, etc...
wwoodscaillon: yeah, that's the intent of the crashdb. the tricky bit is making it so that a user can view their crash report, but nobody else's
warrenWe shouldn't vote on everything, especially non-default things.
wwoodsand fedora devs can view all crash reports, especially those pertaining to stuff they own
caillonwwoods: why?  that seems like too much work
wwoods: let the crashes be public. maybe someone can fix it who isn't a dev
wwoodscaillon: security-sensitive stuff in core files
mclasenwwoods: do you know about similar efforts within the gnome bug team ?
dwmw2f13: hm, ok. I'm ambivalent then.
warrenDo we realistically have enough time and resources to implement this before F8?
c4chrisnotting: agreed
wwoodswarren: it's nearly all the way there, and it's not intended to be enabled by default
f13I'm with notting here, lets not do this as a Feature yet.  lets do it as a Feature when we enble it by default, as /that/ is feature worthy
wwoodssounds reasonable
jeremyagreed
warren-1 Not for a F8 feature, but you may get it working at any time before or after
f13-1
notting-1
dwmw2-1 what f13/notting said
nottingactually, if the *owner* says not-f8, i don't think we need to vote
wwoodsnope
* poelcat moves to CategoryProposedFeature
wwoodswill propose it as a F9 feature once everything's in place
bpepplesounds good.
* c4chris will need to run RSN
dwmw2ok, tickless? Why vote on that? We already did it for i386. We'll do it for x86_64 and PowerPC when they're ready.
didn't we _already_ make noise about tickless?
poelcatNext feature for discussion is: http://fedoraproject.org/wiki/Releases/FeatureTicklessKernel
jeremydwmw2: yes, but we can make more noise!
dwmw2heh
f13dwmw2: because we broke it for i386?
dwmw2oh, did we?
drago01f13: ?
nottingalso, i'd like it tracked so that we know by feature freeze whether it's in or out
b/c i don't want to change it after freeze
f13dwmw2: I'd want some watch over to make sure that we get this in by test2 and not late in the game where it can break again.
notting+1 tickless from mee
f13dwmw2: the fiasco of things like dells not booting with f7, thank tickless (and a dell bug, but nontheless)
dwmw2ah.
f13dwmw2: it also played havoc with the wireless drivers.
bpepple+1, seems like a no-brainer.
dwmw2+1 if it works, -1 if it doesn't. I punt to davej/cebbert
they'll do the right thing
c4chris+1
nottingf13: bah. with wireless any further breakage is just noise
dwmw2and I'll hopefully get powerpc done too some time soon :)
f13yeah, I'm +1 for it, but to be tracked not just a surprise (:
dwmw2: I made my ppc tickless. I turned it off.
dwmw2heh
bpepplehmm, we're out of time, and I believe the rel-eng was planning to have a meeting immediately after us.
nottingshall we run through the rest of these on the mailing list?
caillonand i've got a meeting with the networkmanager guys
f13oh crap, I forgot about that.
go me!
bpepplenotting: +1
dwmw2we punting the kmod thing too then? Would be good to get davej/cebbert to opine on that anyway
c4chrisnotting: +1
abadger1999notting: +1
c4chrisdwmw2: agreed
bpeppledwmw2: yeah, let's do that on the mailing list, too.
nottingactually, which list?
poelcatNext feature for discussion is: http://fedoraproject.org/wiki/Releases/FeatureFixWakeups
bpepplefeatures on the fesco-list.
nottingfesco or -devel?
bpepplejust to keep the noise to a minimum.
nottingpoelcat: you want to send the mails, or i can if you're drowned on the show floor
abadger1999...except there's nothing private about it....
poelcatnotting: please :)
latency is hhigh too
* c4chris bolts cya all later
bpepplelater c4chris
anything else? or should we let the rel-eng start their meeting?
dwmw2_gonelet them start
bpepple-- MARK -- Meeting End
thanks everyone.

Generated by irclog2html.py 2.5 by Marius Gedminas - find it at mg.pov.lt!