From Fedora Project Wiki
fp-wiki>ImportUser (Imported from MoinMoin) |
m (1 revision(s)) |
(No difference)
| |
Latest revision as of 16:26, 24 May 2008
04:06 * | f13 tries to remember what to blog
04:06 < f13> | and of course, gregdek isn't online :/
04:07 <-- | brezhnev has quit ("Download Gaim: http://gaim.sourceforge.net/")
04:13 --> | tibbs (Jason Tibbitts) [n=tibbs@removed] has joined #fedora-summit
04:13 --> | tibbs (Jason Tibbitts) [n=tibbs@removed] has joined #fedora-summit
04:13 --> | tibbs (Jason Tibbitts) [n=tibbs@removed] has joined #fedora-summit
04:13 --> | tibbs (Jason Tibbitts) [n=tibbs@removed] has joined #fedora-summit
04:20 < dgilmore> | f13: everything
04:20 < f13> | hah
04:23 < dgilmore> | f13: buildsys stuff
04:23 < dgilmore> | f13: that i think is a really cool aspect of everything
04:25 < f13> | yeah, there was a long list of things we need, and greg wrote them down. I don't have them on me.
04:25 < f13> | I'm blogging more generic which is OK for now.
04:27 < f13> | dgilmore: where you asking me about subscribing to all Infrastructure/ pages?
04:27 < dgilmore> | f13: yeah i think i got it right
04:28 < f13> | dgilmore: yeah it's Infrastructure/.*
04:28 < f13> | ah! greg put the list in the wiki. Awesome!
04:29 < dgilmore> | hmm i think i did Infrastructure/*
04:30 < f13> | that may work. I just have /.* in my subscription list
04:30 < f13> | er Legacy/.* or infrastructure/.*
04:32 < dgilmore> | i changed to add the .
04:32 * | nirik just pulls the rss feed via liferea... handy that way.
04:47 <@greendiseas> | f13: im about to put up my pictures of stuff
04:47 < f13> | k, I just did a brain flush
04:53 --- | rdieter is now known as rdieter_away
05:09 <-- | llunved has quit (Read error: 110 (Connection timed out))
05:16 --- | rdieter_away is now known as rdieter
05:18 --- | rdieter is now known as rdieter_away
05:33 <-- | codergeek42_ has quit ("Ex-Chat")
05:49 --> | mdomschbot (/msg mdomschbot hello) [n=mdomschb@removed] has joined #fedora-summit
05:58 <-- | mdomschbot has quit (Remote closed the connection)
06:00 --> | mdomschbot (/msg mdomschbot hello) [n=mdomschb@removed] has joined #fedora-summit
06:08 <-- | mdomsch has quit ("Leaving")
06:40 --- | You're now known as thl
07:21 --- | tibbs is now known as tibbs|h
07:35 * | thl wonders if http://spevack.livejournal.com/3986.html needs a s/FAMSO/FESCo/
07:58 <-- | rdieter_away has quit (Read error: 110 (Connection timed out))
08:36 --> | c4chris|w (Christian Iseli) [n=chris@removed] has joined #fedora-summit
11:23 --> | Panzerboy (Unknown) [n=stelian@removed] has joined #fedora-summit
11:39 <-- | pink_trout has quit (Read error: 110 (Connection timed out))
12:39 < jwb_gone> | thl, he might have meant the Fedora Ambassadors project
12:40 < thl> | ohh, do they meet on thursdays these days, too; then never mind
12:46 <-- | paragn has quit (Remote closed the connection)
12:51 --> | paragn (Unknown) [n=paragn@removed] has joined #fedora-summit
13:25 --> | sbarn (Steve Barnhart) [n=steve@removed] has joined #fedora-summit
13:42 --> | llunved (llunved) [n=llunved@removed] has joined #fedora-summit
14:10 < f13> | thl: could go either way. Rex is supposed to bring stuff up to the FESCo
14:11 < thl> | f13, yeah, I've read that
14:11 < thl> | f13, everyone, thx for your work btw
14:23 --- | jwb_gone is now known as jwb
14:24 <-- | llunved has quit (Read error: 110 (Connection timed out))
14:50 --> | Lovechild (David Nielsen) [n=david@removed] has joined #fedora-summit
15:04 --> | rdieter (Rex Dieter) [n=rdieter@removed] has joined #fedora-summit
15:12 < jwb> | f13, around?
15:16 --> | dwmw2_PVG (David Woodhouse) [i=ctrlprox@removed] has joined #fedora-summit
15:42 --> | mdomsch (Matt Domsch) [n=mdomsch@removed] has joined #fedora-summit
15:47 --> | tibbs (Jason L Tibbitts III) [n=tibbs@removed] has joined #fedora-summit
15:50 <-- | halfline has quit (Read error: 104 (Connection reset by peer))
15:54 --> | llunved (llunved) [i=llunved@removed] has joined #fedora-summit
15:55 --> | mspevack (Max Spevack) [n=mspevack@removed] has joined #fedora-summit
15:55 --> | smooge (Stephen J Smoogen) [n=smooge@removed] has joined #fedora-summit
15:55 < mspevack> | good morning
15:55 < smooge> | morning
15:55 --> | markmc (Mark McLoughlin) [i=markmc@removed] has joined #fedora-summit
15:55 < mspevack> | we've got bill, max, greg, and rex here
15:55 < mspevack> | waiting for the others
15:56 < mmcgrath> | mornin
15:56 < jbowes> | good morning
15:57 --> | nasrat (Paul Nasrat) [n=nasrat@removed] has joined #fedora-summit
15:57 < thl> | good {insert local time of time}
15:57 < mspevack> | hi everyone
15:58 < thl> | s/time}/date}/
15:58 < jbowes> | thl: maybe 'hello' would be easier
15:59 < jwb> | mspevack, some questions on the arch stuff have popped up. where should we send those?
15:59 < mspevack> | f-advisory-board, i would say
15:59 < nasrat> | jwb: you mean what does it mean for MY arch ;)
15:59 --> | pink_trout (Pink Ponies) [n=TroutFis@removed] has joined #fedora-summit
15:59 < jwb> | nasrat, yeah but that arch is the only unique one at the moment
15:59 <-- | rdieter has quit (Read error: 113 (No route to host))
16:00 < jwb> | nasrat, because it's transitioning from a primary to a secondary
16:00 < thl> | jbowes, yes, maybe ;)
16:00 < nasrat> | jwb: odd that only sparc seems to be mentioned then
16:01 < jwb> | nasrat, yeah. i think the wiki pages needs to point out more clearly what exactly is going to happen to all the arches. especially rawhide...
16:01 < dwmw2_PVG> | it was suggested that one of the motivations for moving it to secondary was because it was responsible for the slips. https://www.redhat.com/archives/fedora-devel-list/2006-October/msg00199.html and https://www.redhat.com/archives/fedora-devel-list/2006-October/msg00431.html don't seem to support that though
16:02 < dwmw2_PVG> | we work fairly hard to make sure it doesn't lag.
16:03 < jwb> | i'm curious as to where these secondary arch packages are hosted in general. do they become part of rawhide off of fp.org?
16:03 < jwb> | or do they live on some external machine?
16:03 < jwb> | that was discussed, but i don't see it in the wiki page
16:03 * | thl will drive home now; BBL
16:04 < mspevack> | now a few more people in the room:
16:04 < dwmw2_PVG> | Presumably it starts off like Extras -- a second class citizen which nobody really cares about. And then we realise that's not workable when we put our thinking hats on, and try to do something about it, bring it closer.
16:04 < dwmw2_PVG> | as we're sensibly doing with Extrass
16:04 < mspevack> | MaxSpevack, GregDeKoenigsberg, BillNottingham, JackAboutboul, RexDieter, WarrenTogami (who has now left to get donuts)
16:04 < mspevack> | getting closer to quorum
16:04 < mmcgrath> | yum
16:04 < mspevack> | JeremyKatz and DaveJones in the house
16:05 < mspevack> | and JesseKeating
16:05 < smooge> | doughnuts
16:05 < mspevack> | so we're likely to start in a few minutes now
16:05 < smooge> | donuts.. mmmm
16:09 < dwmw2_PVG> | I think we should deal with the arch build process as a separate issue from the set of arches -- change the process by all means but let it settle down and take stock before then changing the set of architectures
16:09 < dwmw2_PVG> | anyway, must sleep.
16:09 --- | dwmw2_PVG is now known as dwmw2_zzz
16:09 --> | gregdek (gaim) [n=gdk@removed] has joined #fedora-summit
16:10 < jwb> | mspevack, can i create a questions section on the wiki page that can be filled in at a later point in time (soon)? sort of like a FAQ?
16:10 < mspevack> | jwb: absolutely, please do
16:10 < jwb> | mspevack, k
16:10 < mspevack> | ok we're starting ================
16:11 < mspevack> | topic the first -- IS/IT, fedora-admin, infrastructure, etc
16:11 --> | rdieter (Rex Dieter) [n=rdieter@removed] has joined #fedora-summit
16:11 < mspevack> | mmcgrath: can you call in?
16:11 < mmcgrath> | yep
16:11 < mmcgrath> | same number as yesterday?
16:11 < mspevack> | abadger1999: can you call too?
16:11 <-- | mdomsch [n=mdomsch@removed] has left #fedora-summit ( "Leaving")
16:11 < mspevack> | dgilmore: how about you?
16:11 --> | warren (Unknown) [n=warren@removed] has joined #fedora-summit
16:12 < f13> | *yawn*
16:12 --> | mdomsch (Matt Domsch) [n=mdomsch@removed] has joined #fedora-summit
16:12 < f13> | jwb: yeah
16:13 < jwb> | f13, see the ArchPolicy wiki page in about 10 minutes. adding some questions that i've been asked frequently
16:13 --> | craigt (Unknown) [n=cthomas@removed] has joined #fedora-summit
16:13 < f13> | jwb: awesome!
16:13 < f13> | did we put on there that ppc is gone?
16:13 < mspevack> | mike -- single sign on is moving along pretty well
16:13 < mspevack> | mike -- some compatability issues
16:13 < jwb> | f13, no. sparc is the only thing explicitly mentioned
16:13 < nasrat> | f13: To ensure that Fedora can be built for any architecture, in a way that doesn't slow down the development for Fedora's Primary Architectures (x86 and x86_64).
16:13 < f13> | ah
16:14 < jwb> | f13, and what "gone" means is one of the questions :)
16:14 < dgilmore> | mspevack: i can make myself available at times
16:14 < f13> | yeah, "no longer a primary arch" as in the official fedora infrastructure will not build for it nor publish it.
16:15 < nasrat> | f13: nice to communicate that to developers who have been working on it
16:15 < jwb> | f13, sure. but it's easier to answer that in my FAQ section. just give me a few minutes
16:15 < mmcgrath> | http://fedoraproject.org/wiki/Infrastructure/AccountSystem2
16:16 < mmcgrath> | and
16:16 < mmcgrath> | http://fedoraproject.org/wiki/Infrastructure/AccountSystem2/Schema
16:16 < nasrat> | f13: so how does that cover say updates, etc to supported. just dumping an arch from the infrastructure without forethought is sucky
16:16 < abadger1999> | mspevack: The conference system tells me the code from yesterday is invalid
16:16 < mspevack> | ok, we're going to discuss is/it priorities
16:16 < mspevack> | bill is on the white board:
16:17 < f13> | nasrat: it was decided yesterday
16:17 < mspevack> | topics
16:17 < nasrat> | f13: so when do I get a mail about it
16:17 < nasrat> | that should have been an action point
16:17 < f13> | nasrat: it is. chill out
16:18 < f13> | nasrat: part of the decision was "ok we want to drop this arch, what all do we need to address to make this happen"
16:18 < f13> | it's not "ok, it's removed from colo/builder configs. Have fun!"
16:18 < nasrat> | f13: that's not clear from the wiki, it's basically just ommited
16:18 < mspevack> | we're starting with the website, fedora.redhat.com
16:18 < mspevack> | How can we kill it>
16:18 < mspevack> | 1) get the logs and see what people are using?
16:18 < jwb> | nasrat, i'm asking those questions in the FAQ. hopefully they'll get answered soon
16:18 < jwb> | nasrat, or turned into action items
16:19 < f13> | nasrat: thanks for pointing that out, we'll fix it.
16:19 < dgilmore> | mspevack: port everything to fedoraproject.org and mirror/ loadbalance it
16:19 < dwmw2_zzz> | so we're going to stop building for _all_ the current rawhide architectures? That's a massively retrograde step for all of them, and _especially_ PPC.
16:19 < dwmw2_zzz> | and in fact for Fedora in general, without even a big-endian machine in the mix to keep developers honest :)
16:20 --> | davidz (David Zeuthen) [i=davidz@removed] has joined #fedora-summit
16:20 < gregdek> | ARCH DECISIONS ARE NOT MADE YET. X86 AND X86_64 ARE CLEARLY PRIMARY. OTHER DISCUSSIONS IN PROGRESS. That is all. :)
16:20 < gregdek> | That's what it says *proposal*.
16:20 < mspevack> | the question is -- what do we want for our web presence?
16:20 < dgilmore> | dwmw2_zzz: ppc and sparc will be thrown in as secondary arches and everything will be built for them. It will be driven by the people wanting those arches
16:20 < nasrat> | gregdek: well it says it's a Policy
16:20 < dwmw2_zzz> | gregdek: ok, thanks for clarification (albeit contradiction of what I thought I saw above about 'is gone'). But I really should be sleeping... :)
16:20 < blizzard> | dwmw2_zzz: go do sleep!
16:21 < mmcgrath> | https://rhn.redhat.com/rhn/help/channel-mgmt/rhn400/en/ch-custom-pkgs.jsp <-- Has fedora.rh.com links ;-)
16:21 < dwmw2_zzz> | blizzard: :)
16:21 * | dwmw2_zzz &
16:21 < jwb> | gregdek, f13, nasrat,, dwmw2_zzz: see http://fedoraproject.org/wiki/FedoraSummit/ArchPolicy
16:21 < jwb> | there's an FAQ section now
16:21 < mspevack> | everything that we did yesterday, and got up on the web is proposals for f-a-b to review, and the community to discuss.
16:21 < mspevack> | it would be best to collect questions up on the wiki like jwb is doing, or on f-a-b list
16:21 < mspevack> | since we want to use this channel today to type about what is *currently* being discussed
16:22 < mspevack> | which right now is our web presence, and what we want it to be
16:22 < dgilmore> | mspevack: i see our web presence as being fedoraproject.org
16:22 * | thl is back on the keyboard now
16:22 < jbowes> | mmcgrath: where should that link go instead? I can get it fixed up
16:22 < mmcgrath> | jbowes: no idea :D
16:22 < dgilmore> | i guess we will need some redirects from anything fedora.redhat.com for awhile to find anything extrernel linking there and have it fixed
16:23 < jbowes> | mmcgrath: somewhere other than a 404, though
16:23 < mspevack> | greg -- what do we need that the wiki doesn't provide us?
16:23 < mspevack> | mmcgrath: it is our most robust site, by far
16:23 < mspevack> | mike -- the fc6 traffic would have killed the wiki completely
16:24 < dgilmore> | i think we need some static pages for rarely changing things
16:24 < dgilmore> | front page should probably be static
16:24 < mspevack> | blizzard -- is our software choice hurting scalability?
16:25 < mspevack> | *crickets*
16:25 --- | mizmo-out is now known as mizmo
16:25 < mspevack> | havoc pennington just walked in
16:25 < dgilmore> | mspevack: our software choice i dont think is hurting it but we dont have scalability right now past one box
16:25 < mspevack> | OH MY GOD, IT'S HAVOC PENNINGTON
16:25 < jbowes> | *applause*
16:25 < mspevack> | ok, we're going to have to temporarily table the website discussion and come back to it later
16:26 < mspevack> | ==========
16:26 < mspevack> | our 10:00 topic was with pete rowley, of OpenID
16:26 < mspevack> | to discuss openID, mugshot, and fedora
16:27 < mspevack> | pete rowley is here, who knows a great amount about openid
16:27 < mspevack> | and donald fischer, who works on mugshot
16:27 < mspevack> | donald -- pete was coming by today to talk to us, do some brainstorming about openID in mugshot.
16:28 < mspevack> | and we've been thinking about using openID or single sign on in fedora
16:28 < mspevack> | greg -- "goal is to lower barrier of entry to fedora. right now lots of systems, license agreement, big headache"
16:29 < mspevack> | greg -- "we're looking for various ways to simplify that and change the way contributions to fedora community work fundamentally
16:29 < mspevack> | we have an LDAP schema for this on the Fedora end
16:29 < mspevack> | ultimately, it's a possibility to allow folks working on Fedora to then also easily participate in other projects that use OpenID
16:29 < mspevack> | greg -- one compelling use case is bugzilla instances talking to each other using OpenID
16:31 < mspevack> | *donald talks about some of the potential gnome project tie-ins*
16:33 < mspevack> | donald -- openID can relieve a lot of the infrastructure issues that exist around letting the Fedora community truly collaborate with each other
16:34 --- | mspevack has changed the topic to: Fedora Summit -- current topic is openID, then IS/IT discussion
16:35 < mspevack> | http://en.wikipedia.org/wiki/Open_id
16:35 --> | corbet (unknown) [n=user@removed] has joined #fedora-summit
16:35 --> | boudiccas (boudicca) [n=boudicca@removed] has joined #fedora-summit
16:35 < blizzard> | http://developer.mugshot.org/wiki/OpenID
16:35 < mspevack> | mike -- with open id integration, should we go full steam ahead with LDAP, or plan how open ID will integrate?
16:36 < mspevack> | pete -- you should keep doing what you're doing, open ID integration should be very easy
16:36 < mspevack> | pete on "what is open id"
16:36 < mspevack> | user-centric identity
16:36 < mspevack> | users are given far more control over their online identity
16:37 < mspevack> | openid is an example of a url-based identity
16:37 < mspevack> | your user id is a URL.
16:37 < mdomsch> | I'd like to hear more about the security of openid
16:37 < mdomsch> | it seems OK for relatively low-value things like blog comments
16:37 < mspevack> | mdomsch: we'll ask about it when he's done with his overview
16:37 < mdomsch> | but I'm not yet convinced for higher-value things
16:38 * | mspevack is not going to try to type all ofpete's open ID description
16:38 < mspevack> | find info on the web and paste it into channel :-)
16:38 < mspevack> | mdomsch: you're on the call, right? I'll throw it to you for a question shortly :-)
16:39 < mspevack> | pete -- open id is secure up to a point
16:39 < mspevack> | major flaw:
16:39 < mspevack> | you're trusting the IDP not to lie to you
16:39 * | mmcgrath has to step away for a minute.
16:39 < mspevack> | IDP == identity provider
16:40 < mdomsch> | mspevack, I'm here, but pete's far from the phone mic
16:40 < mspevack> | pete -- you need to use an authentication scheme that doesn't put your credentials on the wire
16:40 < mspevack> | mdomsch: he's doubled his volume -- i'll just try to type as fast as i can
16:41 < mspevack> | basically you have your IDPs, but you need some level of secure auth on top of them
16:41 < mspevack> | we would need to research some folks who are using OpenID and see what their implementations are
16:41 < mspevack> | dave -- sounds like it would be good for passing stuff around bugzilla, but not for uber-sensitive data
16:41 < mspevack> | bill -- we need some of the profile exchange. for things like bugs, you would want an email address or other information
16:42 < mspevack> | if fedora acts as the identity provider, there's a lot of ways to be secure about the setup
16:42 < mspevack> | pete -- opportunity for Fedora to deploy openID in a secure manner and show everyone else how to do it
16:43 < mspevack> | pete -- certificate-based auth is not very difficult or inconvenient
16:44 < mspevack> | bill -- for what fedora would be using it for, we would almost certainly want to be our own IDP
16:44 * | mmcgrath is back
16:45 < mspevack> | bill -- what sort of things we need stronger auth for -- basically CVS and build system
16:45 --> | mdomsch_ (Matt_Domsch) [n=Matt_Dom@removed] has joined #fedora-summit
16:45 <-- | mdomsch has quit ("Leaving")
16:46 < mspevack> | greg -- we're going to continue down ldap and single sign on anyway, so maybe we use OpenID for some of that. but there might be other places where we need some stronger auth, and openID might not be right for that
16:47 < mspevack> | blizzard -- i'm confused about the role openID plays
16:47 < f13> | jwb: I've tried to answer some of the questions.
16:47 < mspevack> | is it trying to solve the "many accounts" problem? The auth problem? etc
16:47 < mspevack> | pete -- it's trying to solve the single sign on problem
16:47 < mspevack> | blizzard -- let's take this use case. i'm firing off a build from the command line. can the auth still take place using openid?
16:47 < jwb> | f13, k i'll take a look
16:48 <-- | markmc has quit (Read error: 113 (No route to host))
16:49 < mspevack> | *general talk around open ID*
16:51 < mspevack> | donald -- 2 kind of use cases. one is openID for fedora developers
16:52 < mspevack> | another is for fedora users -- when you install Fedora, you can create an openID and all of a sudden you've got connections in user-land between different groups
16:52 < mspevack> | bill -- in firstboot you create your "openid/fedora account" and then later when you want to contribute to Fedora (or some other openID-based project), you're all set
16:54 < mspevack> | greg -- "at what point do we think it's smart to start playing with it? we're going down the path of LDAP based on fedora directory right now"
16:54 < mspevack> | what do you mean by ldap?
16:54 < mspevack> | greg -- we're setting up fedora directory server for people with FP accounts
16:54 < mspevack> | what's on it?
16:54 < mspevack> | mike -- right now nothing, but it will be to replace the current accounting database
16:54 < mspevack> | we have some single sign on stuff, through apache auth to postgres backend
16:54 < mspevack> | we're hoping that LDAP will replace everything in current accounting system
16:55 < mspevack> | main sticking point has been the RHAT bugzilla instance
16:55 < mspevack> | pete -- openid doesn't affect your ldap plans.
16:55 < mspevack> | whatever happens in the future with openid will still be compatible with storing in LDAP
16:55 < mspevack> | greg -- "do we want to try being and openID provider, and in what timeframe"
16:55 < mspevack> | warren -- "sounds like it would just be another interface"
16:55 < mspevack> | everyone talks at once
16:56 --> | jrb (Unknown) [i=jrb@removed] has joined #fedora-summit
16:58 --> | wwoods (Will Woods) [i=wwoods@removed] has joined #fedora-summit
16:58 < mmcgrath> | mspevack: thanks for keeping good notes, its hard to hear whats going on at times.
16:58 < mspevack> | no problem
16:58 < mspevack> | *general chatter right now*
16:59 < mspevack> | about the potential of an openID profile that can tell you "in Fedora, I updated a bugs and did a checkin. In project Foo I did Bar" etc
17:00 < mspevack> | *many jokes about your Fedora OpenID world coming into contact with the "teenage girl" internet world
17:00 < jrb> | mspevack: is it something that other projects would buy into?
17:01 < mdomsch_> | fedora women !
17:01 < mspevack> | jrb: what do you mean by other projects?
17:01 < mspevack> | bill -- from fedora's perspective, what we'd like to do is get SSO/ldap up first
17:01 < blizzard> | girlzz
17:01 < mspevack> | bill -- and then work on adding openid
17:01 < jrb> | mspevack: eg, gnome
17:01 < mspevack> | jrb: donald has mentioned gnome's "echelon" project several times
17:01 < mspevack> | jrb: i'm not familiar with it, but i assume you probably are?
17:02 < jrb> | never heard of it
17:02 < jrb> | oh
17:02 < jrb> | no
17:02 < jrb> | that's something a little different
17:02 < mspevack> | mdomsch_: :)
17:03 < mspevack> | jeremy -- example case
17:03 < jrb> | mspevack: that's more like an uber planet/cia thing: http://live.gnome.org/Boston2006/EchelonForGnome
17:03 < mspevack> | fedora open id -- build system, cvs etc. but you're also a gnome developer
17:03 < jrb> | mspevack: I was more interested in reusing identity for the gnome servers and infrastructure
17:03 < mspevack> | jeremy -- some of the auth mechanisms for both projects are similar
17:03 < mspevack> | jeremy -- and maybe things can be combined
17:03 < mspevack> | jrb -- i think jeremy just hit on your point
17:04 < mspevack> | jrb: but the fact is, if Fedora moves forward with openID, it gives us the chance to convince others to do so as well. that would in theory be part of the benefit
17:05 * | mspevack wants to see the Fedora SSO/LDAP become more of a reality. Whether or not OpenID is the means to that End, or a part of the solution, matters less to me. Making things easier for Fedora contributors is the end goal.
17:05 < mspevack> | that's just my opinion -- but openID does *sound* cool
17:05 < jrb> | mspevack: yeah. We also have this nice smart-card system too...
17:05 < wwoods> | smart cards for everyone! wheee!
17:05 < mspevack> | pete -- an IDP is limited in what it can assert
17:05 < mspevack> | pete -- it can only assert what it has contorl over
17:06 < mspevack> | pete -- if my IDP is scammer.org, i won't be able to assert authority over any URL that is whitehat.org
17:07 < mspevack> | greg -- path forward for now is to finish SSO based on F Directory Server
17:07 < mspevack> | greg -- pete and mike need to get in touch via email, greg is doing that now
17:07 < mspevack> | *sounds like this is wrapping up*
17:07 < mspevack> | and then we'll get back to IS/IT in general for Fedora shortly
17:08 < mspevack> | mike -- question about account system timeline. when do we want it?
17:08 < mspevack> | greg -- what's the critical path for infrastructure?
17:09 < mspevack> | mike -- time and resources
17:10 < blizzard> | "I'd give my left arm to play drums for def lepard"
17:11 < mspevack> | max -- we need to spend the next hour coming up with our infrastructure priorities
17:11 < mspevack> | ===== back at 11:15 ===========
17:11 * | mmcgrath gets notes about what all Infrastructure is up to.
17:11 --- | mspevack has changed the topic to: Fedora Summit -- topic is fedora infrastructure priorities
17:12 < mdomsch_> | mspevack, FYI, one more 1U PE1950 is on its way to PHX to replace the dead hammer box
17:13 --> | graveyard (Ruben Kerkhof) [n=graveyar@removed] has joined #fedora-summit
17:13 < dgilmore> | mspevack: i was going to call in for this part but cant
17:13 --> | glezos (Dimitris Glezos) [n=glezos@removed] has joined #fedora-summit
17:14 < dgilmore> | mspevack: the conference was full
17:16 < mspevack> | who is on the phone?
17:16 * | mdomsch_ coming back...
17:17 < mspevack> | mike, toshio, matt
17:17 --> | green_west (Jack Aboutboul) [n=jack@removed] has joined #fedora-summit
17:18 < green_west> | okay so we are back
17:18 < green_west> | Max: what do we want from a fedora web presence
17:18 < green_west> | Greg: Project Hosting?
17:18 < green_west> | Greg: We need to kill fedora.redhat.com and part of that is figuring out where we are going to move things and to build new things
17:19 < green_west> | Jesse: All the docs are there, so we need to move those somewhere
17:19 < green_west> | Notting: If we kill f.r.c Docs has no static place to put their content
17:20 < green_west> | Max: There needs to be a place where static content lives
17:20 < green_west> | Blizzard: I like the mozilla model, front page has a static, polished layout, the rest of like wiki
17:21 < green_west> | Misc: content managment systems are a royal pain
17:21 < dgilmore> | mspevack: docs.fedoraproject.org would be a good place for docs
17:23 < green_west> | Misc chatter about static content in front of wiki and scaling the wiki
17:23 < green_west> | McGrath: We need a storage backend for the wiki
17:23 < wwoods> | we paid for a professional fp.o redesign, and we have photoshop files laying around unused from that
17:25 < green_west> | Blizzard: there are only a few things that visitors to fp.o initially need, we should target those needs with a static layout
17:25 < green_west> | wwoods: point taken, talking about that
17:25 < iWolf> | Just want to throw my +1 for making the frequently used info static.
17:26 < wwoods> | green_west: I'll again mention that I know (read: am married to) a graphic designer who could work on that for cheap
17:26 < green_west> | Jesse: some pages are static and immutable, those should be static anyway
17:26 < blizzard> | http://www.mozilla.com/en-US/
17:26 < blizzard> | that's the mozilla page
17:26 < green_west> | Max: so we need a two-tiered solution, one with static content on the front few pages, and then a wiki on the back for the rest of out needs
17:27 < green_west> | Blizzard: We need to talk about this during the next Board meeting and just git er done
17:28 < green_west> | Jesse: This isn't stopping us from merging core and extras, why is this a priority
17:28 < green_west> | All: Because it is a high value action
17:29 < green_west> | General Consensus is we need a full time fedora infrastructure person on the red hat payroll
17:29 < mmcgrath> | http://fedoraproject.org/wiki/Infrastructure/Schedule
17:30 < mdomsch_> | mirrors aren't linked directly from front page - links only to download.f.r.c ;-(
17:31 < dgilmore> | mdomsch_: can be fixed
17:32 < green_west> | okay we are looking at the infrastructure schedule and trying to figure out tactical vs. strategic items
17:32 < green_west> | ALL: Random Chatter
17:33 < green_west> | Okay so here are the Line Items of what we have so far:
17:33 < green_west> | 1. Web Site Redesign
17:33 < green_west> | 2. Single Sign On / FDS and later OpenID
17:34 < green_west> | 3. Build System and Packaging DB
17:34 < green_west> | 4. Bugzilla Integration
17:34 < green_west> | 5. Project Hosting
17:34 < green_west> | (these are not prioritized)
17:35 < green_west> | GLUMP!!!!!
17:35 < blizzard> | best word _ever_
17:35 < green_west> | it even auto updates!
17:35 < blizzard> | glumpdates!
17:36 < jrb> | mspevack: any chance of any cross-over with the GNOME infrastructure?
17:36 < green_west> | Okay, so there is discussion about what can be done in parallel to achieve the goals
17:36 < green_west> | Warren: Let's make a dep chart
17:36 < blizzard> | jrb: I suspect the other way around
17:36 < jrb> | blizzard: what do you mean?
17:36 < green_west> | gnome integrates with us!!!
17:37 < green_west> | we are the keymasters
17:37 < jrb> | gnome -> gatekeeper then?
17:37 < blizzard> | jrb: lots of stuff around auth and ldap and build systems, etc
17:37 * | wwoods vinz clortho, keymaster of gozer
17:38 < green_west> | Blizzard: we need to get to a point where we eliminate mail for systems management
17:38 < green_west> | Blizzard: if we develop better information systems, we become more efficient and less annoyed
17:39 < green_west> | Max: Okay, so if we hired someone now, what order of priority do they have
17:39 < green_west> | 1. Build System
17:39 <-- | graveyard has quit ()
17:39 < green_west> | 2. Account System
17:40 < green_west> | 3. Maintainability
17:42 < green_west> | More talk about the website
17:42 < green_west> | Greg: no one ever wants to touch it though, its always someone elses problem
17:43 < green_west> | Jesse: We need one IT person and another webmaster type person
17:44 < glezos> | Note: There are quite a few (new) people in the Infrastructure/Docs teams that would like to help with the new website and tools. They need guidance and probably some core libraries though to work with.
17:45 < green_west> | glezos: yes, we are talking about getting someone to lead this effort
17:45 < green_west> | someone who is paid and accountable
17:45 < green_west> | Greg: We have been talking with seth to figure out what the most accessed pages are and turn them to static content and deploy them in a scalable fashion
17:45 < glezos> | green_west, that would be great and from what I've seen so far, *very* necessary.
17:46 < green_west> | glezos: we are basically trying to justify hiring someone to do this
17:46 < couf> | glezos: +1
17:47 < green_west> | okay so here is the goal
17:47 < green_west> | for website
17:48 < green_west> | Identify Key People and bring them together to rethink/redesign the site
17:48 < green_west> | Blizzard: we need a wikimaster
17:50 < glezos> | green_west, roger that. I think the whole websites-process needs some bootstrapping from an experienced person, ie the building of the core functionality (ACLs, Plone interaction, communication with various teams, decide what content should be put on what tool, etc). After that the community can chip in more easily I think. (eof)
17:51 < green_west> | glezos: yeah, we are trying to figure out what the needs and roles are right now
17:51 < green_west> | Okay, so consensus on website: get a group to rethink site, get a wikimaster, redesign
17:52 < green_west> | Next: Single Sign On
17:52 < green_west> | McGrath: We can get SSO working pretty quickly aside from bugzilla
17:52 < mmcgrath> | Did I say that :D?
17:52 < green_west> | That will be done
17:53 < iWolf> | mmcgrath: :)
17:53 < green_west> | hahahaha
17:53 < green_west> | Next: Maintainability of Infrastructure
17:53 < green_west> | We need hardware
17:54 < green_west> | We need a new rack to put everything in
17:54 < wwoods> | I'm sure gregdek is itchin' to talk about a fedora bugzilla instance... has that already come up?
17:54 < dgilmore> | We need to xenifiy quite a few things to allow for systems to be brought up and taken down as needed
17:54 < green_west> | Jeremy: We need to continue to centralize our hosting
17:55 < green_west> | McGrath: Xen instances for eveything
17:55 < green_west> | Notting: You trust Xen more than I do
17:55 < green_west> | Blizzard: We can look at Amazon SSS
17:56 < green_west> | Jeremy: We need alot of centralized storage in order to maintain a consistent view of the package universe for the build system
17:56 --- | glezos is now known as glezos_afk
17:57 < green_west> | Jesse: mirror admin self service
17:57 --> | graveyard (Ruben Kerkhof) [n=graveyar@removed] has joined #fedora-summit
17:58 * | nirik wonders if anything ever came of bouncer or other better mirror tracking/management.
18:00 < green_west> | Notting: Project Hosting
18:00 < green_west> | First, general hosting space for fedora-* projects
18:01 < green_west> | do we want to provide storage space, bugzilla, other service aside from mailing lists
18:01 < green_west> | Okay, so the consensus is to provide these services for other projects
18:02 < green_west> | we want to centralize information and services as much as possible
18:02 < green_west> | Jesse: what about rolling devel space for other arches?
18:03 < green_west> | people think it will be too much bandwidth for mirroring
18:03 < green_west> | Blizzard: since those download counts are small, why mirror?
18:04 < green_west> | Jeremy: because of locality of download
18:04 < green_west> | Blizzard: other option is for them to have their own hosting
18:05 < green_west> | Jesse: how do those arches spin isos then
18:05 < green_west> | Jesse: via pungi and tools which will be made available
18:06 < green_west> | Jeremy: let's make then do it the same way we do
18:06 < green_west> | PAIN!!!
18:06 * | mdomsch_ drops off the call for a bit
18:06 < green_west> | Warren: we already spoke about many of these action items, and infrastructure has a schedule already
18:07 < green_west> | there other 2 things are content management, which GDK will take responsibility for
18:07 < green_west> | Notting: wait, back to hosting stuff
18:08 < green_west> | Notting: we need to move things off old rh build systems urgently
18:09 < green_west> | Notting: especially all the translation stuff
18:09 < green_west> | Notting: move cvs root to cvs.fedora
18:10 < green_west> | add i18n group to account system
18:10 < green_west> | move accounts from old cvs root, have people move their accounts
18:11 < green_west> | move content
18:11 <-- | Lovechild has quit ("Ex-Chat")
18:13 < green_west> | since this machine currently serves as hosting for tons of this stuff anyway to the outside, we should just move it to the fedora infrastructure
18:14 < green_west> | Notting Still: There are also other features of this system, like granting certain translators permissions to create groups, etc, which need to be translated
18:14 --- | mspevack has changed the topic to: Fedora Summit -- return 1:30
18:14 < mspevack> | ================= lunch, back 1:30 =========================
18:29 --> | wjd (W J Dearborn) [n=wjd@removed] has joined #fedora-summit
18:33 <-- | mether has quit (Remote closed the connection)
18:57 <-- | c4chris has quit (Read error: 110 (Connection timed out))
18:57 --> | c4chris (Christian Iseli) [n=chris@removed] has joined #fedora-summit
18:57 --> | CGI254 ([3ea0658d] CGI:IRC User) [n=3ea0658d@removed] has joined #fedora-summit
18:58 --- | CGI254 is now known as ChitleshGoorah
19:01 --- | glezos_afk is now known as glezos
19:05 <-- | dmalcolm has quit (Read error: 104 (Connection reset by peer))
19:08 --> | dmalcolm (David Malcolm) [i=dmalcolm@removed] has joined #fedora-summit
19:11 <-- | ChitleshGoorah has quit ("CGI:IRC (EOF)")
19:17 --> | ChitleshGoorah ([3ea0658d] chitlesh) [n=3ea0658d@removed] has joined #fedora-summit
19:20 --> | [splinux] (Damien Durand) [n=splinux@removed] has joined #fedora-summit
19:30 < mspevack> | ping
19:30 < mspevack> | i guess i'm still here
19:30 < mspevack> | ok, we're about to get started again
19:31 --- | mspevack has changed the topic to: Fedora Summit -- release methodology (freezes, updates, lifecycle, branding, distro building)
19:33 < mspevack> | skvidal: would you like to call in? we'll be better than we were yesterday
19:33 < mspevack> | mdomsch_: you too
19:33 < mspevack> | get some board members on the phone if we can
19:34 < mdomsch_> | yes, in ~2min, pls start w/o me
19:34 < mspevack> | ok == STARTING NOW ==
19:34 < mspevack> | Jesse has the floor
19:34 < mspevack> | mspevack will continue to be IRC bitch
19:34 < mspevack> | gregdek will continue to be main note taking guy
19:35 < gregdek> | thx :)
19:35 < dgilmore> | mspevack: i know some in extars wont like it but I feel we have to stop the rolling release for everything but the development tree
19:36 < mspevack> | gregdek: if you didn't do such a good job, we'd stop asking :-)
19:36 < dgilmore> | mspevack: donttell him that :D
19:36 < mspevack> | dgilmore: jesse seems to agree
19:37 < mspevack> | jeremy -- this goes back to our platform discussion yesterday
19:37 < mdomsch_> | mspevack, concall is at capacity; can't get in
19:37 --> | markmc (Mark McLoughlin) [i=markmc@removed] has joined #fedora-summit
19:37 < mspevack> | mdomsch_: i'll bump it up again
19:37 < mspevack> | i upped it to 7 people
19:38 < mdomsch_> | nope
19:38 < mspevack> | bill -- what are we trying to define? How we define and build something that we release at release time?
19:38 < mspevack> | greg -- "what does an initial release of Fedora version Foo look like, and how is it maintained"
19:38 < mspevack> | jesse -- "what do we as FP want to make as our initial spin?"
19:38 < mspevack> | jesse -- it won't be the size of what Core 6 was
19:39 < mspevack> | mdomsch_: it's up to 10 people
19:39 * | mmcgrath still prefers the idea of core and smaller 'modular' sub projects.
19:39 < wwoods> | FP = Fedora Platform?
19:39 < mspevack> | greg -- single platform or multiple? greg thinks multiple with a common base
19:39 < mspevack> | wwoods: in that case, FP == Fedora project
19:39 < mmcgrath> | but that never really worked in reality.
19:39 < mspevack> | mmcgrath: pls expand
19:39 --> | mether (Rahul Sundaram) [n=ask@removed] has joined #fedora-summit
19:40 < mspevack> | phone -- "as non RH person, like the idea of Core being a small initial set of stuff, with other pieces that can be placed on top.
19:41 < mspevack> | bill -- as Fedora Project being spun by Jesse, does he want to make it possible for others to spin stuff and call it Fedora?
19:41 < thl> | dgilmore, i agree with " i know some in extars wont like it but I feel we have to stop the rolling release for everything but the development tree", nervertheless there seems to be a strong interest in a "rolling release" model -- could we provide both somehow?
19:41 < dgilmore> | that being gcc, kernel and glibc? and things needed by those?
19:41 < mspevack> | jesse -- we want people to be able to make spins of packages in the FU signed by fedora keys and call it fedora
19:41 <-- | gregdek has quit (Read error: 104 (Connection reset by peer))
19:41 < mspevack> | jesse -- part of that is making sure they have a fedora standard base, etc.
19:42 < mspevack> | Fedora Standard Base is a subset of the "Fedora Platform"
19:42 < dgilmore> | thl: they could if they follow development
19:42 < mspevack> | bill -- should everything on fp.o's fedora 7 torrent be bulit by Jesse?
19:42 < mspevack> | jesse -- "i don't think that's necessary"
19:42 < mspevack> | jeremy -- "i don't think that scales"
19:42 < dgilmore> | thl: we would need a freeze at release time to get iso's etc done
19:43 < mspevack> | blizzard -- says something very similar
19:43 < dgilmore> | after that it would be updates
19:43 < thl> | dgilmore, many people won't touch it as long as we tell them "might eat babys"
19:43 < mspevack> | 1) Desktop Gnome
19:43 < dgilmore> | thl: true and at times it eats babies and other things.
19:43 < mspevack> | 2) LAMP
19:43 < mspevack> | 3) Java?
19:43 < dgilmore> | Desktop KDE
19:44 < mspevack> | *nod*
19:44 < dgilmore> | Development
19:44 < mspevack> | dgilmore: make it easy for people to put together their own recipies, whatever it is
19:44 < dgilmore> | mspevack: sure
19:44 < mspevack> | blizzard -- desktop environment with Java stack, eclipse, jboss. here's your Fedora Java developer workstation
19:45 < skvidal> | hey, cool - now we can get those java rpm bindings going
19:45 < dgilmore> | mspevack: we dont have the infrastructure for on demand iso's
19:45 < jbowes> | skvidal: +1
19:45 < skvidal> | jbowes: haha
19:45 < mdomsch_> | dgilmore, once pungi exists, that'll help
19:46 < skvidal> | jbowes: I forgot to close the </sarcasm> tag
19:46 < jbowes> | skvidal: you joke, but i'd use em!
19:46 < skvidal> | jbowes: you're deeply broken
19:46 < jbowes> | skvidal: jum is coming. march '07
19:46 < skvidal> | jbowes: I dare you
19:46 < skvidal> | :)
19:46 < dgilmore> | mdomsch_: yeah what would be cool is a web interface i select what i want and get an email with a download link once an iso is created fro me
19:46 < mspevack> | bill -- the question is, if we are going down this road with "random" spins, do we remove the concept of self-hosting?
19:46 < mdomsch_> | not all spins need to be self-hosting
19:46 < mspevack> | jesse -- I think so. We must. As long as anaconda can reach the entire "Fedora universe"
19:47 < warren> | Fedora Everything Blue-Ray!
19:47 < mdomsch_> | soon and very soon
19:47 < mspevack> | Fedora Source RPM spin
19:47 < mmcgrath> | Blue-Ray orbital laser
19:48 < warren> | mmcgrath, you may buy only 80k of them
19:48 < jbowes> | skvidal: well, probably more like April 1 '07
19:48 < skvidal> | jbowes: yes
19:49 < dgilmore> | i would say we want to provide a defult set of an installable tree
19:49 < dgilmore> | say a server install and a desktop install
19:49 < mdomsch_> | dgilmore, but derivatives may delete and add components from rest of Universe as they need
19:50 < dgilmore> | mdomsch_: sure they can.
19:50 < dgilmore> | mdomsch_: some people will just want a desktop install what we provide is good enough
19:50 < mspevack> | jesse -- on demand ISOs is further in the future.
19:51 < dgilmore> | anything else in the universe can be installed latter thats what yum and pirut are for
19:51 < warren> | dgilmore, we need to build the ability to create that easily first. Then it would be relatively easy to implement the on demand tool.
19:51 < dgilmore> | warren: i agree
19:51 < mspevack> | jesse -- whatever tool we use to create our release will be publicly available, so anyone else can make a release and we can host torrents.
19:51 < mspevack> | but that's not quite yet on-demand isos.
19:52 < dgilmore> | we could use the most commonally selected components to make up some defualt offerings
19:53 < mspevack> | *sidetracked into a livecd debate*
19:54 < nirik> | would it be ok to call it Fedora X if all the packages assembled were from fedora buildsys and not if someone adds other packages to it? Just a thought for naming...
19:54 < abadger1999> | livecd note: Just before FC6, I was able to create a working livecd with kadischi but not with pilgrim. That could have changed with FC6 updates.
19:55 < abadger1999> | thl: I think rolling releases are going to become more important to people. Especially if we start pegging our releases to a single component (like GNOME).
19:55 < mspevack> | bill -- who is accountable for deciding what goes into the "Gnome Desktop"
19:56 < abadger1999> | If XFCE is consistently 1.5 months after the GNOME release, for instance, the xfce developers will want to get that into the post-release stable tree.
19:56 < mspevack> | bill -- for example
19:56 < dgilmore> | mspevack: Gnome project
19:56 < dgilmore> | mspevack: what gnome provides as their desktop we provide as their desktop
19:56 < mspevack> | gregdek -- desktop and server
19:57 < mspevack> | dgilmore -- we mean "the package manifest that is the "Fedora Desktop (Gnome)" build, for example.
19:57 < nirik> | abadger1999: Xfce is unfortunately not very consistant about their release dates. :) It's ready when it's ready... but that only adds to your point I think. ;)
19:57 < thl> | abadger1999, agreed, but on the other hand some people want not that much updates as they bear the risk of breakage and are big in download; thus we should try to gie something to those that want the "classical model of a stable release" and another solution ("altrenatives") that is not that risky as devel, but still rolling release
19:57 < dgilmore> | mspevack: it would be whatever is needed by the gnome desktop as provided by gnome. nothing more nothing less
19:59 < dgilmore> | a server image would have all the server things. postfix sendmail exim, mysql postgresql. which could be guided by a Fedora Universe steeing committe
19:59 * | dgilmore needs to learn to spell
20:00 < mspevack> | gregdek -- central case is having a spin at the time the initial release goes gold. We've got "Fedora 7" and that consists of all the packages that are "Gold" in time. Fedora base, Desktop, server, and a lot of extras. All of that goes into some set of CDs.
20:01 < dgilmore> | yep
20:01 <-- | Panzerboy has quit (Remote closed the connection)
20:01 < mdomsch_> | dgilmore, are a lot of folks disliking the "rolling extras" model?
20:01 < mdomsch_> | that epel won't take care of?
20:01 < dgilmore> | mdomsch_: there is quite a few who want to keep it
20:02 < mdomsch_> | given how often FC is released, I don't personally have a problem with rolling extras
20:02 < dgilmore> | mdomsch_: i personally feel we cant
20:02 < dgilmore> | mdomsch_: there will be no more extras
20:02 < dgilmore> | there will be only Fedora
20:02 < mspevack> | bill -- pungi needs to divorce stage 1, stage 2, boot.img
20:02 < mspevack> | from the tree
20:02 < mdomsch_> | one could argue an intermediate updates-testing stage would be nice
20:03 < mspevack> | mdomsch_: seth is getting off if you want to try to get on
20:03 < skvidal> | mdomsch_: enjoy
20:03 < dgilmore> | mdomsch_: it would be beneficial
20:03 < skvidal> | mspevack: say bye to everyone for me
20:03 * | mdomsch_ is on
20:03 < mspevack> | jesse -- we all agree that we will have a platform, and the FSB will be a part of that.
20:03 < mspevack> | FSB is the least number of packages you can have and call it Fedora
20:06 --> | Panzerboy (Stelian Iancu) [n=stelian@removed] has joined #fedora-summit
20:06 < abadger1999> | dgilmroe: We're still going to release a F.Platform out of the Fedora Universe. We could release a F.Desktop, F.Core, F.Whatever made up of a selected group of packages that want to stabilize for a release while the rest of F.Universe rolls on. (Just one alternative)
20:07 < abadger1999> | Sorry, s/dgilmroe/dgilmore/.
20:07 < mspevack> | branding debate -- are we using the term Fedora as a weapon to enforce quality control or not
20:07 < ChitleshGoo> | F.Kde (for example) too ?
20:07 < dgilmore> | abadger1999: sure we can
20:08 < quaid> | fedora-as-branding-weapon++ :)
20:08 < ChitleshGoo> | +1 fedora-as-branding-weapon
20:09 < nirik> | if all the packages used are "offically built" Fedora packages, shouldn't you be able to call any collection of them Fedora? Or you mean get approval to do so?
20:10 < ChitleshGoo> | nirik: once approved, i guess it can be called fedora, (correct me if i'm wrong)
20:10 < mspevack> | that's what there is some argument about, nirik
20:11 < mspevack> | what if someone makes a pile of Fedora packages, but it doesn't boot, or etc.
20:11 < mspevack> | these are some of the complaints
20:11 < nirik> | ok, just wanting to clarify... branding as a weapon could mean lots of things. ;)
20:11 * | mmcgrath likes where this discussion is headed...
20:11 < ChitleshGoo> | +1 mspevack e.g geda packages
20:11 < mspevack> | greg -- "the brand experience of Fedora is taking whatever you want from the Fedora Universe and calling it Fedora"
20:11 < dgilmore> | mspevack: if its created with our super install tree creation tool it should boot and be called fedora
20:12 < dgilmore> | i use my lame i want an image tool then no
20:13 * | mmcgrath ponders a webcast :D
20:14 < mspevack> | gregdek -- need 3 simple rules
20:15 < mspevack> | 1) "official" Fedora spins
20:15 < mspevack> | nevermind...
20:15 < mspevack> | this conversation is nearly impossible to summarize on IRC
20:15 < mspevack> | it's getting very philosphical
20:15 < mdomsch_> | defer decision of who is blessed to the Board
20:15 * | nirik guesses the 3 simple rules aren't all that simple. ;)
20:16 < mspevack> | the idea of the FSB is basically being killed right now
20:16 < mmcgrath> | heh
20:16 < mspevack> | and what is gaining momentum is "here's the blessed Fedora builds"
20:16 < mspevack> | that maybe the Board has endorsed
20:16 < mspevack> | there's the "Fedora Remixes" that folks in the community are building, maybe we never hear about it, etc.
20:17 < mspevack> | and then there's the stuff that "cannot ever be called Fedora" which is the person who takes Fedora packages, adds in livna, and spins an iso.
20:17 < mdomsch_> | orbital lazer control included in FSB?
20:17 < mspevack> | that's an attempt at summary
20:18 < mspevack> | and the fedora remixes are likely to bubble up and get blessed
20:20 < mmcgrath> | mspevack: whats the agenda for the rest of the day?
20:20 < mspevack> | 5 minute break
20:20 < mspevack> | next is:
20:20 < quaid> | that seems like a nice model
20:20 < mspevack> | freezes, lifecycle, updates
20:21 < mspevack> | what we just covered was "distro building, flavors, a bit on governance"
20:22 < dgilmore> | mspevack: governece by FUSCo :D
20:23 < thl> | dgilmore, how about FUDSCo ? (no, I don't know what the "d" could stand for)
20:24 * | thl afk now
20:24 < thl> | hae a nice summit
20:24 * | wwoods forms an anti-Fedora Universe splinter group, Stop The Fedora Universe
20:24 < mspevack> | Fedora Universe Comittee
20:24 <-- | warren has quit (Read error: 104 (Connection reset by peer))
20:24 * | mspevack coughs
20:24 < mmcgrath> | heh
20:24 --- | You're now known as thl_afk
20:25 < quaid> | oh, yeah, now that is acronym I can really get behind
20:25 < mether> | quaid: fedora universe and documentation steering committee?
20:26 < quaid> | FUCSo I mean
20:26 < quaid> | Fedora Universe Committee for Steering and Organizing
20:27 < quaid> | Fedora Universe Committe Unite!
20:27 < quaid> | FUCU
20:29 < mspevack> | ok, back to order ====================
20:29 < mspevack> | Jesse still has the floor
20:29 < mspevack> | freezes, lifecycle (Legacy), updates
20:30 < mspevack> | lmacken: can you get on the phone?
20:30 <@greendiseas> | he is on the phone
20:30 < mspevack> | k
20:30 < lmacken> | yep yep
20:31 < mspevack> | .... and here we go.
20:32 < mspevack> | jesse -- freezes updates, lifecycles, and rawhide. oh my!
20:32 < mspevack> | how do we manage freezes
20:32 < mspevack> | what do updates look like in a rolling/not rolling universe
20:32 < mspevack> | what's lifecycle look like?
20:32 < mspevack> | what's rawhide look like?
20:32 < mspevack> | jesse -- currently we gather everything that's in core collections
20:32 < mspevack> | we drop it in directory
20:32 < mspevack> | we run some stuff
20:32 --> | warren (Unknown) [n=warren@removed] has joined #fedora-summit
20:32 < mspevack> | we sync it to the world
20:33 < mspevack> | it takes a long time, just with our package set
20:33 < mspevack> | it's got lots of problems
20:33 < mspevack> | the user experience pretty much sucks, but because rawhide is a snapshot in time, so be it.
20:33 < mspevack> | ananconda doesn't always work, so you can't always do a direct rawhide install
20:33 < mspevack> | rawhide spin is nothing more than create repo call
20:34 < mspevack> | jeremy
20:34 < mspevack> | rawhide is a collection of packages
20:34 < mspevack> | installer is divorced from packaged set as much as humanly possible -- right now pretty close
20:34 < mspevack> | packages go out at whatever frequency
20:34 < mspevack> | daily installer images
20:34 < mspevack> | you point them at a tree
20:36 < mspevack> | GOAL -- divorce installer from package set
20:36 <-- | wilee has quit (Read error: 104 (Connection reset by peer))
20:36 < mspevack> | you can use a known good installer against any package set
20:36 < mspevack> | rawhide is just a collection of packages
20:36 < mspevack> | and we provide a daily installer image "basically a rawhide installer" that you point at a tree
20:37 < mspevack> | rawhide == "all the latest built packages in Fedora Universe"
20:37 < mspevack> | rawhide will be pushed at fixed times
20:37 < mspevack> | with ability to do an emergency push
20:37 < warren> | yay!
20:38 < mspevack> | and then there will be a rawhide installer
20:39 < mspevack> | rawhide is a rolling development release
20:41 < wwoods> | pushed 'at fixed times' - what sort of frequency? Weekly? Monthly? I assume it'd have to pass a few cursory install checks before it would get pushed?
20:41 < mspevack> | wwoods: daily, i think
20:41 * | mspevack asks
20:41 < wwoods> | ah, okay
20:42 < mspevack> | jesse -- "fixed times" means "at least daily, possibly faster"
20:42 <-- | craigt has quit ("Leaving")
20:43 < warren> | jesse -- In the past extras was built against yesterday's rawhide (sometimes problematic)
20:43 < warren> | Now everything will be built in the same place.
20:43 < warren> | (me: yay!!!)
20:43 < mspevack> | open question -- signing of packages
20:44 < mmcgrath> | good question.
20:44 < mspevack> | jesse -- is everyone happy with the future of rawhide?
20:44 < lmacken> | signing server
20:44 < mspevack> | issue with signing server -- not necessarily every piece of it is OSS
20:44 < mspevack> | someone mentioned nCipher
20:45 < dgilmore> | mspevack: serial console access with a small handful of people able to sign. perhaps 3 or 4
20:45 < mmcgrath> | mspevack: If we do go the 'respin' route. Who's going to sign those?
20:45 < mspevack> | keep packages around "for a while" that have gone into rawhide
20:45 < mspevack> | you mean, custom people who build an ISO out of the FU?
20:46 < mspevack> | all those packages will already be signed
20:46 < warren> | Keep packages around "for a while", especially the ones that went public.
20:46 < mspevack> | warren: i already wrote that :-)
20:46 < warren> | oh
20:46 < mmcgrath> | k
20:47 < nirik> | would be nice if there were signers spread out in diffrent timezones and enough to where waiting on them wasn't a bottleneck...
20:47 < mspevack> | nirik: in theory signing will be automated
20:47 < mspevack> | if you can trust everything that goes into the build system
20:47 < nirik> | ah, cool.
20:47 < mspevack> | then you can have a signing server
20:47 < mmcgrath> | hmm
20:47 < mspevack> | but if you can't trust everything that goes into the build system, you can't.
20:47 < warren> | nirik, signing server allows abstraction of the master signing from the developers.
20:47 < mspevack> | isn't that the whole point of the signing server?
20:48 < mspevack> | NEW TOPIC -- export control
20:48 < warren> | nirik, theory: developers login to (something), authenticate with their personal (something) key, which then signs using the real key.
20:48 < mspevack> | Fedora has to inform the legal team:
20:48 < mspevack> | 1) we're making a release
20:48 < mspevack> | 2) we're not adding any new crypto stuff
20:48 < mspevack> | 3) they make a filing
20:48 < mspevack> | bill -- "#2 will hurt!"
20:49 < nirik> | warren: that woud be great... then developer would control when a package was signed and released?
20:49 < warren> | nirik, and announced with the update, yes.
20:50 < warren> | nirik, details we still have to figure out, and implementation...
20:50 < nirik> | cool.
20:50 < mspevack> | so, there's an action item here for legal team, which max owns)
20:50 < mspevack> | NEW TOPIC -- freezes
20:50 < mspevack> | one of the reasons why we are so adamant about being able to tag packages and have collection inheritance and get stuff out of buildsys based on a collection tag is so that we can have sane freezes
20:51 < warren> | (freezes necessary to do a release)
20:51 < mspevack> | internall -- we have a collection, say dist-fc6. when you build, it's built against that and then tagged for it. so on so forth. that's UNFROZEN
20:51 < mspevack> | you know what, i don't think I need to type all the "here's how we've always done stuff" stuff
20:52 < mspevack> | *jesse speaks about how we do builds*
20:52 < warren> | *Current way of doing things, manually moving packages during a freeze, works only for a small set of people.*
20:53 < mspevack> | packages are tagged into collections
20:54 < mspevack> | trees are built based on a collection that all has the same tag
20:55 < warren> | *New way: Do not freeze entire thing at once, people do not stop creating due to a freeze*
20:56 < warren> | *(Optionally things after this point can be pulled in if they solve specific problems identified by (someone)*
20:56 < mspevack> | warren: can you take over with the notes? :_)
20:57 < warren> | *(This is a theoretical and experimental process that we have to refine.0
20:57 < warren> | mspevack, ok
20:57 < mspevack> | warren: thank you :-)
20:58 < warren> | * Theoretical debate going on whether this is a good idea or not. davej gave the example of stable/devel kernel branches that became only one branch in 2.6.)
20:58 < warren> | * Jesse gestured and pointed at blank whiteboard in an attempt to will his mind upon it.
21:00 < warren> | * (Jesse is writing on board)
21:01 < warren> | Jeremy -- We require a way to take a snapshot in time, allow things to move forward, and cherry pick things to bring into that snapshot.
21:02 < warren> | * (jeremy talks about painful ways we did things in the past)
21:02 < warren> | * (when six people were working on the distro)
21:02 < warren> | * notting repressed those past painful memories.
21:03 < warren> | * WHITEBOARD
21:03 < warren> | Composes happen from collection tags.
21:03 < warren> | Freeze tags are created for composes.
21:03 < warren> | Fixes are cherry-picked from base collection tag (i.e. dist-fc7).
21:03 < warren> | Base tag stays rolling (devel)
21:03 --> | craigt (Unknown) [n=cthomas@removed] has joined #fedora-summit
21:04 < sbarn> | is everything thats current been discussed available on the wiki in one place at the moment?
21:04 < sbarn> | or that was discussed i should say so far
21:04 < warren> | sbarn, not really, this is relatively new within RH itself
21:05 < sbarn> | ok
21:06 < warren> | * (Talking about test composes)
21:07 < mspevack> | example:
21:07 < mspevack> | dist-fc7 tag applies to every package built from the devel branch currently.
21:07 < quaid> | sbarn: they've been posting the IRC logs and minutes at the end of the day, as well as folks summarizing on planet.fedoraproject.org
21:07 < mspevack> | Test1 freeze is announced. Date hits, we create fc7-test1 tag (freeze tag)
21:07 < mspevack> | we tag the latest dist-fc7 packages with this freeze tag
21:07 < mspevack> | We create a test repo of all packages tagged "fc7-test1". QA happens on these packages, or some subset
21:07 < sbarn> | quaid, i just saw, that, wasnt like that yesterday thanks
21:08 < mspevack> | realistically a lot of these packages will never get looked at.
21:08 < mspevack> | Fixes are applied to the devel branch and built. Resulting specific package builds get tagged with "fc7-test1" at the discretion of the release team
21:08 < mspevack> | at final release, when we announce the freeze, we create the fc-7 branch
21:09 < mspevack> | devel begins to be tagged with the "fedora 8" tag
21:11 < dgilmore> | mspevack: seems sane
21:12 < mspevack> | it's all directly out of Jesse's mouth
21:12 < mspevack> | it all needs to get run by FESCO
21:12 --> | vpv (Ville-Pekka Vainio) [i=vpivaini@removed] has joined #fedora-summit
21:12 < mspevack> | next -- UPDATES AND LIFECYCLE (including Fedora Legacy)
21:13 --- | mspevack has changed the topic to: Fedora Summit -- updaes/lifecycle, Fedora Legacy
21:13 < dgilmore> | mspevack: perhaps at say test3 point in time we tag fedora8 for devel and have a test3 updates tree with bugfixes for fedora7
21:14 < mspevack> | probably -- that wasn't exhaustive, it was just what I could type :-)
21:14 < mspevack> | bill -- updates and how you handle a released Universe
21:15 <-- | craigt has quit (Remote closed the connection)
21:15 < mspevack> | new step involved to get an updated package out for a released product
21:15 < mspevack> | 1) build package from branch
21:16 < mspevack> | 2) request update using tool
21:16 --> | craigt (Unknown) [n=cthomas@removed] has joined #fedora-summit
21:20 < lmacken> | i can hardly hear whoever is talking (jeremy)?
21:20 < dgilmore> | mspevack: push tools would do the rest>
21:20 < dgilmore> | ?
21:21 --> | mizmo_ (Máirín Duffy) [i=duffy@removed] has joined #fedora-summit
21:21 <-- | mizmo has quit (Read error: 110 (Connection timed out))
21:22 < mspevack> | wwoods: can you get on the call?
21:22 < mspevack> | rex -- is updates-testing something that we want
21:23 < mspevack> | rex -- make updates-testing part of the build root
21:23 < mspevack> | *general confusion*
21:23 < mspevack> | this feels like a topic that will need more discussion
21:24 * | glezos loves the *atmosphere* logging :)
21:24 --- | stickster_work is now known as stickster
21:24 < wwoods> | mspevack: sure, I don't have the conference code though
21:24 < mspevack> | we are going to put will on the spot and ask him about updates-testing
21:24 < mspevack> | will:
21:25 < mspevack> | "we talked about policy for doing updates in a recent meeting"
21:25 < nirik> | nice features for the update tool: marking of security/bugfix, flagging to only release when dev gives final ack, etc... ;)
21:25 < mspevack> | "basically one of the problems is that we don't have enough people to enforce any of this, but what we'd like to have happen is that if at all possible, someone should install the package and get results"
21:25 < mspevack> | luke -- "arch maintainers (from yesterday) can be required to give acks?"
21:25 < mspevack> | will -- "that might work"
21:26 < lmacken> | http://fedoraproject.org/wiki/Infrastructure/UpdatesSystem ("ideal" workflow i came up with)
21:32 < mspevack> | significant conversation, which will best be summarized by other later. it's impossible to do it justice in channel, i'm sorry
21:34 * | nirik would also like to see the rss feed per package/branch... would be cool to subscribe to packages you care about and get updates fed to you...
21:34 < mspevack> | and for the record, davej sucks. because the kernel sucks :)
21:34 < warren> | * Someone proposed a Love It/Hate It system like used in Mugshot for easy and quick user feedback for test updates.
21:34 < mspevack> | i think that was jeremy
21:35 <-- | ChitleshGoorah has quit ("CGI:IRC (EOF)")
21:35 < blizzard> | jeremy is a h8r
21:35 < jwb> | mspevack, for the record, that should be the kernel developers suck. Fedora's mantra is "UPSTREAM"
21:35 * | nirik would also like to see a popcon like package for fedora, but I suppose I will have to write it myself to get anything moving there.
21:35 < mether> | nirik: do it then?
21:36 < warren> | * greg: How do these tools prevent gratuitous updates?
21:36 < nirik> | in my copious free time? :)
21:37 < warren> | * (talking about how to do automated QA in the buildsys)
21:37 < warren> | * (post build process tests in the buildsys)
21:37 < warren> | * Take care of the "OMG it is breaking the world" case.
21:37 < warren> | * Build package in branch, do post-build processing, get thumbs up
21:38 <-- | llunved has quit (Read error: 104 (Connection reset by peer))
21:38 < warren> | * Request update using tool, optional flag at build time to put it into the buildroot for building another thing that relies upon it.
21:39 < warren> | * Fill out, Why is this being updated? Is this security? Bug numbers this update is closing? (for possible automatic closing) Click button to push update.
21:39 < warren> | * Behind the scenes, release team: somebody wants to push update, please sign & move & push
21:41 < warren> | * Repo management (like pup and pirut) is coming, allowing users to enable updates-testing and do other things.
21:41 < warren> | * Jesse is the repo-monster.
21:42 < warren> | * Updates go to updates-testing
21:42 < warren> | * We have haters, or lovers
21:42 < warren> | * Timeout occurs, nagmail happens to maintainer
21:42 < warren> | * "timeout, please check for haters, check for lovers, go to this tool to request pushing update"
21:42 < warren> | * davej -- Great for developer morale. "No one likes you."
21:42 < dgilmore> | works for me
21:43 < warren> | * Greg -- Get hate/love integrated into mugshot.
21:43 < warren> | * Go to tool, click "push this as a final update"
21:43 < dgilmore> | no one likes me either so davej and I can can sit in the corner
21:43 < warren> | * davej -- package consumed with hate, either push (unjustified hate), or destroy. It is up to the maintainer.
21:44 < warren> | Maybe write "Extra Justification" in the box.
21:44 < warren> | * davej -- I want a hate-o-meter.
21:44 < dgilmore> | heh
21:44 < wwoods> | "Waived - 'Don't hate the player, hate the game'"
21:45 < warren> | * jeremy -- We are not in terrible shape without this. We are not worse off than we are today. This would improve the process, and it is definitely something that the testing team looking into would make sense...
21:45 < warren> | * jesse -- hatebot!
21:45 < blizzard> | <davej> even the bots hate me
21:45 < wwoods> | h8b0t?
21:45 < warren> | * jeremy -- friends that are nerds
21:45 * | warren loves quotes out of context.
21:47 < warren> | * Packagers duty to check the love/hate
21:47 < wwoods> | lmacken: as a side note, I was just starting to write a TurboGears-based thing for QA feedback on updates-testing packages
21:47 < wwoods> | so this is rad
21:47 < wwoods> | but I can't get it to run :/
21:47 < lmacken> | wwoods: awesome!
21:47 < lmacken> | where is the code ?
21:48 < lmacken> | i'm working on getting test6.phx.fedora to be our 'TurboGears application server'
21:48 < nirik> | a popcon like package could also indicate how many people installed/used a updates-testing package... might be of use...
21:48 < wwoods> | lmacken: nowhere. I started it this morning and threw it away when I realized how your tool worked
21:48 < wwoods> | heh
21:48 < lmacken> | haha
21:48 < warren> | * Signing Server enables update team (who approves pushes) are not master signers.
21:49 < warren> | * Proposal is for ALL packages to go through this update process. This must be proposed and discussed by FESCO.
21:49 < warren> | * Technical concern - canceling update
21:49 < wwoods> | pysqlite2.dbapi2.OperationalError: no such table: package
21:49 < warren> | * Not a big problem, because not in buildroot unless you explicitly ask.
21:49 < lmacken> | <code>tg-admin sql create</code> ?
21:50 < warren> | So... big news of this is *ALL* updates in Fedora in an existing release will have update announcements.
21:50 < lmacken> | weeeee
21:51 * | warren after all committees approve it.
21:51 < lmacken> | and useful metadata for pup/pirut
21:51 < warren> | lmacken, will you be available during Thursday's FESCO meeting to help discuss the tools support angle of this?
21:51 * | dgilmore has been wanting that
21:51 < wwoods> | lmacken: *forehead slap*
21:51 < warren> | lmacken, I think FESCO's acceptance of this will rely on understanding how nice the process is with your update processing system.
21:51 < lmacken> | warren: i think i can make it
21:52 < lmacken> | i have to write 2 papers and take 2 more finals this week
21:52 < warren> | lmacken, putting something online, even just screenshots to show ideas, would help for that meeting.
21:52 < lmacken> | warren: http://fedoraproject.org/wiki/Infrastructure/UpdatesSystem
21:52 < lmacken> | warren: like that ?
21:52 < warren> | lmacken, do you want a publictest* machine to run a mock-up of this?
21:52 < lmacken> | warren: i'm already working on test6.phx, but a publictest machine would probably be better
21:52 < warren> | lmacken, damn, you rock!
21:52 < lmacken> | i know.
21:53 < lmacken> | :)
21:53 < wwoods> | Hooray! Let's all talk to zoidberg!
21:53 < wwoods> | I mean, lmacken!
21:53 < dgilmore> | lmacken: i can change test6 to publictest6 if you want
21:53 < warren> | lmacken, just change the IP address to a publictest address and reboot the xen guest, and update the InfrastructurePrivate page.
21:53 < warren> | or dgilmore
21:53 < dgilmore> | lmacken: or warrens idea
21:53 < lmacken> | dgilmore: sure, althought i haven't put any firewall rules on it yet
21:53 < lmacken> | i still need to hack up a pyroman config profile for our xen guests
21:54 < dgilmore> | lmacken: public ports are 22 80 443 8887 8888 8889
21:54 * | mmcgrath notes the infrastructure team as a whole is worried about port 22 being open.
21:54 <-- | glezos has quit (Remote closed the connection)
21:54 < dgilmore> | lmacken: denyhosts needs putting on also
21:54 < dgilmore> | mmcgrath: yup sure am
21:54 < warren> | Yes, that is why infrastructure team needs to approve anything that runs on publictest* addresses.
21:54 < warren> | IDEA
21:54 < warren> | 1) Approval required of infrastructure team
21:55 < warren> | 2) denyhosts must be configured
21:55 < warren> | 3) ssh key authentication ONLY
21:55 < lmacken> | +1
21:55 * | dgilmore is ok with that
21:55 * | iWolf nods
21:55 < mmcgrath> | warren: +1 obviously but its going to be difficult to enforce since most people will have root on that box.
21:55 < warren> | Hmm, we have infrastructure quorum now, let's ratify? =)
21:56 < warren> | hmmm
21:56 < mmcgrath> | s/that box/their box/
21:57 < warren> | mmcgrath, you're right, that is a flaw in this.
21:58 < mmcgrath> | I suppose due diligence is required on our part. It won't be perfect but its something.
21:58 < dgilmore> | warren: end users wont have root only the people managing the test service
21:58 < warren> | dgilmore, right
21:58 < warren> | ok
21:58 < warren> | Ratify?
21:58 < mmcgrath> | ratify
21:58 < dgilmore> | +1 from me
21:58 < warren> | +1
21:58 < mmcgrath> | +1
21:58 < iWolf> | +1
21:58 < warren> | Given we don't have formal voting rules or a defined quorum size, I guess this is ratified.
21:59 < iWolf> | :)
21:59 < lmacken> | good good
22:00 * | warren posts new policy to list.
22:00 < warren> | lmacken, go ahead and change the IP and reboot the guest, but update the docs to reflect it.
22:00 < lmacken> | warren: k
22:06 < warren> | MEETING IS BACK!
22:07 < warren> | * We have concluded that we want Borat to be the speaker at the next FUDCON.
22:07 < warren> | Or Colbert
22:08 < mmcgrath> | Making of the OSS free for people!
22:08 * | dgilmore votes for Borat
22:08 * | mmcgrath votes for Colbert in a mustache.
22:09 < warren> | * Legacy becomes Fedora Security Team
22:09 < warren> | * Core becomes Fedora (something), fully external
22:09 < warren> | * Fedora Security Team first Tracks problems and pokes maintainers to fix things.
22:09 < warren> | * Fedora Security Team may also fix things.
22:10 < warren> | * jeremy -- proposes "fedora sustainability team"
22:10 < warren> | * other people don't like the name
22:11 < jwb> | renaming something isn't going to really change anything
22:11 < warren> | * warren -- Proposed: If an old distribution fails to be maintained (older, less popular), retire it.
22:11 < warren> | * people don't like that idea, because it is difficult to measure
22:11 < jwb> | if you don't fix the issues Legacy has right now, you could call whatever you want and it still wouldn't work
22:12 < warren> | * jeremy -- New Proposal: Security updates happen longer (in some way they are discussing)
22:12 < warren> | jwb, one idea is that Updates being fully in the open will allow lowering the bar and making it easier to sustain an older distro.
22:13 < warren> | * mspevack -- I wish you waited a little longer before it becomes legacy. (What he hears often.)
22:13 < jwb> | warren, "fully in the open" meaning no embargo?
22:13 < warren> | * jeremy -- Make it two instead of one.
22:13 < warren> | "running up the flag pole"
22:14 < warren> | * Possibly distro being fully in the open would allow community to help RH engineering to create updates, thereby reducing pressure on RH engineering.
22:14 < warren> | * (talking about when to stop bug fixes on a FC release)
22:15 < warren> | * jeremy -- Proposed: Stop bug fixes and security updates for FC7 at FC9+1 month.
22:15 --> | llunved (llunved) [i=llunved@removed] has joined #fedora-summit
22:15 < jwb> | warren, perhaps but i doubt it. legacy releases are fully in the open now once they become legacy and it hasn't fixed things
22:16 < wwoods> | So.. that would lengthen the support window by ~60 days?
22:16 < warren> | * Debating whether or not Fedora will allow Legacy to happen on Fedora infrastructure or not.
22:17 < warren> | jwb, we don't know for certain
22:17 < warren> | * jeremy -- Clarity of distro retirement is very important to end users.
22:18 < smooge> | Fedora DeadRelease Team
22:18 < warren> | There is skepticism here whether *anybody* can realistically maintain an older distro.
22:18 < jwb> | warren, the only thing that is certain is death and taxes ;)
22:19 < warren> | * jeremy -- Big thing, reducing scope.
22:19 < smooge> | To be honest.. I have to agree.. at this point, I just try to figure out a way to get someone converted to the nearest Centos release.
22:19 < warren> | * jesse -- currently legacy maintains two releases
22:20 < warren> | * Fedora Security Team: tracking things for everything during lifecycle
22:20 < warren> | * FC7 -> FC9+1 month
22:21 < warren> | * Honesty is important here regarding Legacy's ability to work or fail.
22:21 < warren> | * jesse -- Legacy went from many to few people
22:22 < warren> | smooge, "jeremy -- If you need longer term maintenance, then you should look at longer term options like CentOS or RHEL."
22:22 < warren> | * rdieter -- Compromising, extending lifetime
22:23 < warren> | * notting -- reality, legacy has dropped down to one packager and a few other people
22:23 < warren> | (jkeating said too)
22:23 < warren> | * davej -- upgrade from FC8 to FC10, will be testing that?
22:23 < warren> | * everyone -- yes
22:23 < smooge> | we used to have to test upgrade paths from 3 releases back
22:24 < warren> | NOTE: Things discussed here are proposals, nothing decided yet.
22:24 < smooge> | RHEL-4.x/RHEL-3.0.3 -> 5.x
22:24 < smooge> | understood by me
22:25 < smooge> | most of the people I know who have Fedora Legacy issues had it installed at a colo and needs updates
22:25 < wwoods> | I think FC(n-3)->FCn upgrades should still be tested, and I think upgrades should get a *lot* of work
22:25 < warren> | * jesse -- By making these adjustements, this obsoleting the need for an official Fedora Legacy project.
22:25 * | warren this point not entirely in agreement here yet.
22:27 < warren> | * Extending useful lifetime to 13+ months while stopping to pretend that legacy was useful. We need truthful messaging.
22:27 < warren> | (Greg)
22:29 < warren> | * Change 13 month policy for FC5 immediately?
22:29 < warren> | * Yes, we think, need to check on this.
22:30 <-- | dmalcolm has quit (Client Quit)
22:30 < jwb> | 13 months from initial release?
22:30 < jwb> | or 13 months after going legacy?
22:30 < mdomsch_> | jwb, 13 from initial
22:30 < warren> | jwb, example: FC7 lives on until FC9 + 1 month
22:30 < warren> | 13 months is an estimate
22:30 <-- | sbarn has quit (Remote closed the connection)
22:31 * | corbet thinks this is a big step in the right direction
22:31 --> | dmalcolm (David Malcolm) [i=dmalcolm@removed] has joined #fedora-summit
22:31 < wwoods> | It seems like time spent maintaining (e.g.) FC3 would be better spent working on making upgrades from FC3 -> FC6 worked well
22:31 < warren> | * Strong perception "legacy is not a success"... because "legacy is not a success"
22:31 < warren> | wwoods, yes, this is subject to more discussion.
22:31 < wwoods> | warren: right, just my opinion
22:31 < smooge> | Strong perception that legacy is dead
22:33 < mdomsch_> | at 13 months, don't nuke the old tree from the servers, just don't post new updates
22:33 < warren> | * wwoods keeps his hate strong.
22:33 < nirik> | would help to have a tested/approved update method using yum... that would allow more folks to update instead of being stuck on old releases.
22:33 < lmacken> | warren: publictest2 is now the TurboGears guest (xen2)
22:33 < mdomsch_> | so people running can upgrade when *they* want, not when we release FC(N+2) on our scheudle
22:34 < warren> | nirik, please propose to FPB later to discuss and possibly ratify that as a standard.
22:34 < warren> | * Legacy team and RH engineering must approve of the discussed plan.
22:34 < warren> | nirik, actually, please begin on FAB
22:34 < warren> | nirik, are you on FAB?
22:35 < nirik> | nope... I'm not on anything. ;)
22:35 < warren> | nirik, please talk to me after and we'll fix that.
22:35 < wwoods> | nirik: upgrading a running system is kind of hard, so yum updates will always be a bit dicey
22:35 < mdomsch_> | wwoods, I've gone 3->4->5->6 on mine :-)
22:35 < wwoods> | but I personally feel that yum upgrades should be fully supported
22:35 < mdomsch_> | not without some pain, but got there
22:35 < wwoods> | mdomsch_: same here.. and it's kinda gross
22:35 < nirik> | yeah, although I have done it with the last few releases and the problems have gone down a lot.
22:35 < warren> | wwoods, many yum upgrade bugs are worth investigating, might not be possible to totally fix, but it is worth doing as an exercise.
22:36 < wwoods> | e.g. had to nuke my beagle conf to get it to work in FC6
22:36 < mdomsch_> | once I nuked 3rd party repo packages first
22:36 < wwoods> | mdomsch_: oh lord, yes. also that.
22:36 < mdomsch_> | upgraded, then put them back
22:36 < nirik> | see the decrease in issues on http://fedoraproject.org/wiki/YumUpgradeFaq for example
22:37 <-- | tibbs has quit ("Konversation terminated!")
22:37 < wwoods> | mdomsch_: the dirty secret is that anaconda just silently ignores 3rd-party packages. tada! no conflicts!
22:38 < mdomsch_> | wow
22:39 < warren> | * jeremy -- reason why we don't officially support upgrade in place, is to allow the flexibility to make fundamental system changes. For example: static dev to udev.
22:39 < warren> | * max -- Makes it an incredibly hard sell, (Fedora vs. other distro)
22:40 < warren> | * jeremy -- upgrades have been failing badly for Ubuntu
22:40 < wwoods> | right, it would be nice if there was a yum-based upgrade that worked *almost* like that
22:40 < warren> | Like I said, we should *TRY* to make it work, and test it, and fix bugs when we can.
22:40 < wwoods> | i.e. it pulls all the packages, reboots into an updater image, runs the update, and then boots the new system
22:40 < warren> | But it might not be possible to guarantee it to work.
22:41 < wwoods> | someone could probably write a simple python script to do that, in fact.
22:41 < warren> | Fixing bugs found during in-place yum upgrades often do fix legitimate problems that would happen in anaconda upgrades.
22:42 < warren> | * Greg - "Put some pants on by the way!"
22:42 < warren> | * Max - "Let's all understand the way the QA was done for FC6, and what level of community help we had, your time being used for (talking too fast..."
22:43 --- | warren has changed the topic to: Fedora Summit -- wwoods (QA test lead) interrogation
22:43 < f13> | ==== Taking a moment to talk to wwoods. ====
22:43 * | f13 is taking over as IRC monkey
22:43 * | warren hands banana to f13.
22:44 <-- | [splinux] has quit (Read error: 54 (Connection reset by peer))
22:45 < f13> | Question, did bug zappers really work well?
22:45 < dgilmore> | f13: for maybe a week
22:45 < f13> | Jack: went pretty well when I did it, but fell off when I went back to school. Lack of bug day coordination.
22:45 < f13> | Jack: people doing it felt unloved due to lack of coordination and communication.
22:45 < f13> | Jeremy: Works for gnome, but has a constant driver
22:46 < warren> | *Talking about ranking system with scores and ranks.
22:46 < warren> | "level 1 packager with +3 sword."
22:47 < f13> | Max: Unless the people whom you're asking to do bug work are seeing quick results, the incentive goes away
22:47 < f13> | Completely right ^^^
22:47 < f13> | Max: there seemed to be no control over blocker bugs
22:48 < f13> | Max: Did Will have the power to block the release?
22:48 < f13> | Answer: yes.
22:48 < warren> | * talking about FC6Blocker but not being empty, because anybody could add
22:48 < warren> | * Freedom to add is important
22:48 < wwoods> | http://fedoraproject.org/wiki/QA/ReleaseCriteria
22:48 < warren> | * Release cabal decided what were the "actual real blockers" and try to be happy at release time
22:49 < warren> | * release cabal did stop release a few times to fix critical bugs.
22:49 < warren> | * Max: File a bug, block "test 1 triage", and team goes through it.
22:49 < warren> | * Jesse -- Maintainers should look at bugs
22:49 < warren> | * notting -- maintainers use blockers as a TODO list
22:50 < warren> | * jeremy -- from prior experience, it is a full time job to do triage
22:50 < warren> | * davej -- there will always be nasty bugs that can't be fixed (TOO MANY!)
22:50 < warren> | * max -- Few people that bother to do good things in bugzilla tend to be ignored.
22:51 < warren> | engineers of packages in core, make what appear to be arbitrary decisions that frustrates and disincentives community involvement in QA.
22:51 < ajax> | disincents.
22:51 < warren> | * greg -- fundamental intractable problem that is QA, identify most useful activities ... (talking too fast)
22:51 < warren> | * greg -- bug day: virtual equivalent of buying people a pizza to do work
22:52 < warren> | * promise to people around them to be there (compared to 3 hour party in evercrack)
22:53 < warren> | * We need commitment from engineering^WFedora contributors to meet community half-way, so QA people feel good about doing it.
22:53 < f13> | talk about testing.108, and RHTS type stuff.
22:53 < f13> | the RHTS guys are very excited about bringing it to Fedora
22:55 < warren> | (RHTS is an automated suite of regression tests.)
22:55 < f13> | Not all parts are open, and there is chicken / egg problem. No tests to build infrastructure around, no infrastructure to build tests.
22:59 <-- | graveyard has quit ()
23:00 < f13> | Discussion regarding watching bug reports and escelating to "common" issues.
23:01 < warren> | * Greg - People do Extras because they need it. People do translations (even with shit infrastructure) and it succeeds because they need it.
23:02 < warren> | * Greg - Motivations for communicaty projects are not intrinsic, (digging to bugzilla shit)
23:02 < smooge> | does this cover about people with large number of NEW bug events to their name from multiple releases?
23:02 < f13> | Manager/SIG escelation there.. :/
23:02 < warren> | Greg: how do we concentrate on making sure that people give a shit about this?
23:03 < warren> | wwoods -- better tools
23:03 < warren> | wwoods -- add incentives
23:03 < warren> | jesse -- enable people to fix shit themselves
23:03 < warren> | greg -- motivation for filing bug, dependent on increased expectation that bug will actually be fixed
23:03 < warren> | solve motivation problem?
23:05 < smooge> | well as a customer and talking to other people.. you don't get new bugs when you see that xyz maintainer hasnt changed a Bugzilla from NEW to TAKEN (or whatever) since 1999
23:05 < warren> | warren said:
23:05 < warren> | People driven QA requires constant manual input
23:05 < warren> | "Bug day" is one-time and did not create something self-sustaining.
23:05 < warren> | Better investment is to create automated ranking systems (like GNOME Bugzilla) that encourages fundamental changes in volunteer behavior, due to a feeling of competition and accomplishment.
23:05 <-- | markmc has quit ("Leaving")
23:06 < smooge> | it is also used as sales part from other companies. "Well if they are so good, why isnt NetworkManager bugs ever cleared"
23:07 < warren> | * Discussing pros and cons of using Bugzilla for a scoring system.
23:09 < warren> | * Possibility of adding scoring/ranking/reward system to upstream bugzilla, because this is something that projects need in general.
23:10 < mdomsch_> | split the problem into multiple parts
23:10 < mdomsch_> | it's not "all bugs" but "class of bugs #1, #2, #3, ..."
23:10 < warren> | * jeremy -- Bugzilla is featuritis overload. Everyone that uses Bugzilla is forking it.
23:10 < mdomsch_> | like we did with the BuidRequires
23:11 < mdomsch_> | karma
23:12 < mdomsch_> | http://fedoraproject.org/wiki/QA/FixBuildRequires tried
23:12 < mdomsch_> | but the folks who did it really didn't do it for karma points
23:13 < mdomsch_> | they did it for the love of a "better" distro
23:13 < wwoods> | mdomsch_: yes, that was the interesting thing to me
23:14 < f13> | speaking of, wwoods, we need to get those karma points redeemed soon.
23:15 < wwoods> | f13: indeed
23:17 < warren> | * Possibly scoring could be done by an async process that queries the Bugzilla and updates scores elsewhere periodically.
23:17 < warren> | * wwoods talking about RHTS ...
23:19 < warren> | * writing test plans not exciting, but helpful
23:20 --- | jwb is now known as jwb_gone
23:22 < warren> | * warren: Incentivize writing documentation with Love/Hate links, which contributes to the point ranking system!
23:22 < warren> | * Single Sign-On is where the points are tracked!
23:22 < warren> | * Race: Elvish
23:23 < smooge> | maybe you could tie it into Mugshot?
23:24 < warren> | smooge, yes, there was discussion of this.
23:24 < warren> | smooge, Mugshot creates excellent opportunities in devel information sharing
23:25 < warren> | * (Add some kind of points tracking thing to the account system schema maybe)
23:26 < dmalcolm> | wwoods: BTW, I've been working on code that makes it very easy to create a regression test for a package's rpmlint output within RHTS
23:26 < wwoods> | dmalcolm: yay!
23:26 < warren> | * Points are not about earning stuff, but earning community respect and "cred"
23:26 < smooge> | I mean I am more interested in recommending bugs/software than songs
23:27 < warren> | smooge, as Havoc said it earlier today, "If you don't have any nerd friends, you wont see nerd things."
23:27 < f13> | mdomsch_: thanks for that update (:
23:27 < warren> | smooge, Mugshot shows you stuff depending on who your friends are.
23:27 < smooge> | hmmm I guess I don't have any friends :)
23:28 <-- | mspevack has quit ("Leaving")
23:29 < mdomsch_> | np
23:29 * | mdomsch_ nuked the 16GB of trees a few days ago to reclaim the space
23:29 < f13> | heh
23:30 < wwoods> | http://fedoraproject.org/wiki/Testing/Meetings/20061109
23:31 < dgilmore> | smooge: you and me both :D
23:31 < green_west> | quaid: ping
23:31 < green_west> | quaid: are you on the call
23:33 < warren> | green_west, the bridge is dead now
23:34 < green_west> | yes, but if he was on, he wouldnt be disconnected
23:34 < warren> | green_west, new calls incoming after 5pm stopped
23:40 < warren> | * jeremy -- Meritocracy, people writing code are those who make decisions about the direction of Fedora.
23:41 < mmcgrath> | s/writing code/doing works/
23:41 < mmcgrath> | :P
23:42 < warren> | * Greg -- What about Live CD?
23:42 < warren> | * notting -- LiveCD does not rely on anaconda
23:42 < warren> | * jeremy -- Problem: skill levels required to do a certain thing, VERY DIFFICULT to hire someone to work on anaconda.
23:44 < warren> | * max -- wanting to solve problem of RH engineering not being able to focus on key priorities important for Fedora
23:44 < warren> | * jeremy -- difficult o_O
23:44 < jrb> | hrm
23:44 < warren> | * jeremy -- Fedora does not do a good job of enumerating what is important
23:44 < warren> | thus the goal of this summit
23:44 < smooge> | how do people who actually want to use Fedora but aren't coders (but say admins/IT/bugtesters) have a say in the direction of Fedora
23:45 < warren> | smooge, meritocracy... if you aren't contributing in a big way, it is difficult to be respected in a meritocracy.
23:45 < warren> | smooge, if someone without skills but resources (like money), they could put up bug bounties
23:45 < warren> | Also... if someone without TIME but money, they could put up bug bounties.
23:46 < warren> | smooge, also... respect alone isn't useful in getting something. WORKING CODE can be.
23:47 < warren> | (Still talking about resource scarcity)
23:47 < jrb> | warren: is the goal of the summit to enumerate what's important to fedora?
23:47 * | mmcgrath wonders how other OSS projects deal with resouce issues.
23:48 < warren> | * jeremy -- Going to world where everything is external, everyone has to be held to the same standards. (Community contributing code determines direction and what happens. ... maybe...)
23:48 < warren> | jrb, one of many goals of the summit
23:50 < warren> | jrb, nothing really was decided in finality here. This was mainly a valuable exercise in putting heads together, posing questions and thinking about options.
23:50 < jrb> | warren: do I get to vote? (-;
23:50 < smooge> | warren, I understand that but that goes to the question: why would I want to use Fedora for some project if I have no standing in finding that a current direction is broke?
23:51 < warren> | jrb, in the actual decisions that result from recommendations of this summit, yes.
23:52 < warren> | jrb, have time to stop by tomorrow on the last day of the summit? Topics ... LiveCD, What's in Fedora? "non-free" repos/links, FUDCon/Summit, Max's ideas
23:52 < jrb> | warren: alright...
23:52 < warren> | * jeremy -- We work upstream, we work upstream, we work upstream.
23:54 < warren> | * GENERALLY WHAT THIS DISCUSSION IS ABOUT: Resource scarcity problem. RH can't do everything that is important for Fedora. Fedora is becoming fully external. Meritocracy is important to Fedora. Hopefully this will allow Fedora itself to easier add things that are important to Fedora.
23:55 --- | wwoods has changed the topic to: Fedora Summit -- Resource scarcity problem - how do we decide what to work on?
23:55 < warren> | * notting -- (ask eng managers) What things we would like in Fedora we can have egineers work on upstream?
23:55 * | wwoods no longer being interrogated!
23:55 --- | mizmo_ is now known as mizmo-out
23:56 < mdomsch_> | laptops
23:56 < dgilmore> | mdomsch_: they are useful :D
23:56 < warren> | * notting -- interesting thing, why nobody in RHEL cares about LiveCD?
23:56 * | mdomsch_ thinks laptop features should be a key driver for FC7
23:56 < spot> | because there is so little corporate Linux on desktop?
23:56 < warren> | * Priority: LiveCD or Xen?
23:56 --- | tibbs|h is now known as tibbs
23:56 < mmcgrath> | notting: I bet they'll care more when xen, virtualization, and stateless linux takes off.
23:56 < warren> | something loses =(
23:56 < mdomsch_> | lowering power consumption
23:57 < mdomsch_> | wifi / bluetooth
23:57 * | mdomsch_ is hardware-centric
23:57 < warren> | * Fundamental issue that we try to do as much of everything in upstream in possible
23:57 < dgilmore> | warren: id choose xen over livecd
23:58 < warren> | dgilmore, thus it is the job of Fedora to beat^Wencourage a community project to seriously make LiveCD happen.
23:58 * | mdomsch_ needs to go cook steaks soon
23:58 < mmcgrath> | dgilmore: but you could do both ;-)
23:58 < dgilmore> | mmcgrath: sure i could
23:58 * | dgilmore is not that sadistic
23:58 < mmcgrath> | heh
23:58 < mdomsch_> | part of LiveCD challenge has been - no offense jeremy
23:59 < mdomsch_> | jeremy saying "sure, I'll work on it, when I've got time"
23:59 < warren> | * jesse -- Fedora goes after upstream to fix things (crap makefiles, HIG compliance, etc.)
23:59 < mdomsch_> | which scares others away, because they know he'll do a good job
23:59 < mdomsch_> | if he gets the time
23:59 < warren> | * jesse -- Do not patch locally, go after upstream to fix it for everyone.
23:59 < warren> | * should be job of maintainer
23:59 < mdomsch_> | e.g. the "don't bite off more than you can chew"
--- Day changed Mi Nov 15 2006
00:00 < warren> | * Max - Let's go drink.
00:00 < warren> | * Jeremy - Let's sum up what we did today.
00:00 < wwoods> | It is totally beer-thirty, you guys
00:00 < wwoods> | GTFO
00:00 < mmcgrath> | hah!
00:00 * | mdomsch_ is gone
00:01 <-- | rdieter has quit (Remote closed the connection)
00:01 * | wwoods stops lurking on the call and heads home