[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

FESCo Meeting Minutes for 2006-08-17



= 2006 August 17 FESCo =

Meeting Summaries are posted on the wiki at:

  http://www.fedoraproject.org/wiki/Extras/SteeringCommittee/Meetings

== Attending ==

 * warren
 * thl
 * bpepple
 * c4chris
 * rdieter
 * tibbs
 * abadger1999
 * dgilmore

== Summary ==

=== Mass Rebuild ===
 * http://www.fedoraproject.org/wiki/Extras/Schedule/FC6MassRebuild
 * Packagers get 3 weeks to rebuild just like AWOL policy.
 * Packages that haven't been rebuilt get orphaned and the packages
won't ship in the FC6 repository.
 * AWOL package process started for ignacio as he has a bunch of
packages that others might have to pick up as dependencies.
   
=== comps.xml ===
 * c4chris sent out nagmails.  Many people have updated their packaged
in comps.xml.in.
 * There are preliminary plans by the Packaging Committee to remove the
group tag in the spec file and only have the information in comps.xml.
 * c4chris thinks the only way to make 100% accurate nagmails is to
record a group for all packages, which could be an invisible group for
some packages.  This information might be appropriate for the Package
Database.
 * c4chris and bpepple will start a comps SIG.
 * Even command-line tools should be in comps.
 
=== Legacy in buildroots ===
 * Waiting on legacy being able to access the buildsystem so they can
build ppc updates.
  
=== Ctrl-C Problem ===
 * Infrastructure says figuring out a solution before the new VCS is
probably too much work.
 * New VCS prototyping is waiting on two new servers to be installed at
the end of August.  Hoping to have something ready for Extras people to
test by FC7t1.
 
=== Packaging Committee Report ===
 * Started discussing deprecating and removing Group tag.  No timeline
yet but definitely a post-FC6 change.
 * ldconfig wording change to clarify the meaning.
 * Talked about changing the meeting day/time so there's time to send
the committee report to FESCo via email.
 
=== Sponsorship Nominations ===
 * dgilmore and c4chris accepted as sponsors.
 * New rule, nominate new FESCo members for sponsorship if they are not
already; debate about upgrading them will be per normal.
 * Qualifications to be a sponsor needs to be discussed.
 
=== Misc ===
 * New Package Review template approved:

https://www.redhat.com/archives/fedora-extras-list/2006-August/msg00361.html
 * wiki user setup to send changes to FESCo list when /Extras/Schedule*
is changed.
 * kmod discussion
  * Packaging Committee will discuss the technical aspects, whether to
allow them in Fedora Extras/Core at all is being passed up to FAB.
 
=== Free discussion ===
 * Maintainer Responsibility::
http://fedoraproject.org/wiki/JasonTibbitts/MaintainerResponsibilityPolicy
  * tibbs solicits input on the draft policy.  There are several open
questions that need answering therein.

== Log ==
{{{
(09:55:23) ***warren here.
(09:55:32) warren: mdomsch, you there?
(09:55:48) mdomsch: warren, yes
(09:57:40) ***jima pops some popcorn for the show
(09:59:05) ***cweyl settles in to lurk...  silly people at work who
presume to schedule their meetings during FESCo!
(10:00:16) thl has changed the topic to: FESCo meeting in progress
(10:00:22) thl: hy everyone
(10:00:26) thl: who's around?
(10:00:36) ***bpepple is here.
(10:00:36) ***c4chris_ is here
(10:00:55) c4chris_ is now known as c4chris
(10:01:28) rdieter: here
(10:01:35) tibbs: I'm here.
(10:01:38) ***cweyl is lurking (rabble)
(10:01:41) thl: okay, so let's start slowly
(10:01:56) thl has changed the topic to: FESCo meeting in progress --
M{ae}ss-Rebuild
(10:01:57) ***abadger1999 here
(10:02:16) thl: scop not around
(10:02:26) thl: do we want to discuss this further
(10:02:37) thl: open issues: How much time need people to rebuild their
stuff? What happens with packages that haven't been rebuilt or marked as
not needing a rebuild by deadline X?
(10:03:04) c4chris: what's the delay in the AWOL policy?
(10:03:15) ***dgilmore is here
(10:03:29) c4chris: I think we should use the same delay
(10:03:35) dgilmore: thl: no need for discussion  Lets just do it
(10:03:59) tibbs: The delay is three weeks and three days, BTW.
(10:04:00) thl: c4chris, I don#t know the exact delay from AWOL
(10:04:27) c4chris: tibbs, sounds about right
(10:04:32) thl: let's say: give people three weeks to rebuild their
stuff?
(10:04:39) c4chris: thl, +1
(10:04:40) bpepple: thl: +1
(10:04:46) thl: that leaves some time to fix the remaining stuff before
FC6 ships
(10:04:47) tibbs: +1
(10:04:50) warren: +1
(10:05:02) dgilmore: +1
(10:05:05) rdieter: +1
(10:05:11) thl: okay, three weeks
(10:05:27) thl:  What happens with packages that haven't been rebuilt or
marked as not needing a rebuild by deadline X?
(10:05:39) tibbs: Orphaned?
(10:05:47) c4chris: Let's deal with the not rebuilt package when we have
an idea how many there are
(10:06:11) tibbs: We should know in advance so that people won't just
expect the current package to rull into FC6.
(10:06:23) tibbs: s/rull/roll/
(10:06:27) c4chris: mmh
(10:06:31) c4chris: ok
(10:06:37) dgilmore: im kindof scared if ignacio  doesnt step up  he has
alot of packages
(10:06:43) thl: dgilmore, +1
(10:06:43) c4chris: orphaned sounds about right
(10:06:55) thl: jwb, you wanted to contact ignacio iirc
(10:07:07) bpepple: dgilmore: Should other people be brought in to help
him?
(10:07:10) jwb: thl, i sent him an email.  no repsonse
(10:07:19) thl: jwb, thx
(10:07:39) tibbs: Packages not rebuilt certainly shouldn't get
autobranched to FC6, but I don't think they should be deleted.
(10:07:50) dgilmore: i sent him one about 6-8 weeks ago  asked how
things were going  got nothing
(10:07:55) c4chris: we never delete packages
(10:08:00) c4chris: just orphan them
(10:08:10) thl: tibbs, well, that would mean that someone had to
maintain them through the entry lifecicle of FC6
(10:08:15) thl: we need to delete them
(10:08:46) dgilmore: delete binaries  but not cvs
(10:08:55) c4chris: dgilmore, oh right
(10:09:35) tibbs: The binaries should not get out to FC6 unless they've
been rebuilt.
(10:09:57) thl: okay
(10:10:02) c4chris: so the threat is: they won't ship in FC6 until
rebuilt (or a short explanation why they are not rebuilt...)
(10:10:16) jwb: that seems reasonable
(10:10:19) rdieter: yup
(10:10:26) thl: so just to be sure: +1 for "delete all packages that are
not rebuild in time" please
(10:10:27) tibbs: +1
(10:10:34) jwb: +1
(10:10:34) c4chris: +1
(10:10:36) abadger1999: +1
(10:10:38) rdieter: +1
(10:10:42) dgilmore: +1
(10:10:51) tibbs: +1
(10:10:58) abadger1999: For ignacio specifically, has the AWOL packagers
process been started?
(10:11:26) dgilmore: abadger1999: no  but it really needs to
(10:11:32) jwb: agreed
(10:11:41) bpepple: dgilmore: +1
(10:11:48) warren: Want me to attempt to use non-email to contact him?
(10:11:52) BobJensen: dgilmore: +1
(10:11:57) warren: Ask him what he wants us to do with his packages.
(10:12:00) abadger1999: warren: +1
(10:12:05) thl: warren, maybe that would be a good idea
(10:12:08) c4chris: warren, +1
(10:12:09) Daveman: :o
(10:12:15) tibbs: If someone has his phone number, it would certainly be
a good idea to try it.
(10:12:21) BobJensen: I tried calling all I got was VM
(10:12:36) Daveman: FC6?
(10:12:36) dgilmore: warren: +1
(10:12:42) warren: I'll give it a try.
(10:12:58) thl: k -- lets ignore the question "remove pacakges from cvs
or only orphan them later"
(10:13:09) thl: maybe we could use a scheme like this:
(10:13:25) thl: remove all packages that were not rebuild three weeks
before FC6
(10:13:53) thl: and remove all those from cvs when the branch for FC6
happens, that don#t have a new maintainer yet
(10:14:37) thl: so let's move on
(10:14:47) dgilmore: thl: they stay in cvs  but get flagged so they wont
build
(10:15:08) thl: dgilmore, could work, too
(10:15:24) thl: how hard it restoring of deleted things?
(10:15:43) thl: (in cvs)
(10:15:48) c4chris: easy
(10:15:50) dgilmore: harder  than fixing a spec file set to not build
(10:16:04) thl: let's stop here
(10:16:13) thl: we can discuss this later 
(10:16:16) thl has changed the topic to: FESCo meeting in progress --
Use comps.xml properly
(10:16:22) thl: c4chris, the nag mails worked great
(10:16:29) c4chris: thl, yup
(10:16:33) jwb: yes, they do
(10:16:45) thl: dgilmore: automate comps file during push or via cron
(10:16:49) c4chris: Now I have a few questions...
(10:16:51) thl: dgilmore, did you look into this?
(10:16:53) abadger1999: c4chris: I don't think I received any though
(and I should have)
(10:17:07) thl: ohh, shoot (dgilmore, we'll discuss this later)
(10:17:08) c4chris: abadger1999, I'll check
(10:17:12) dgilmore: thl: i was going to help who was  but i cant rember
who that was. 
(10:17:36) thl: dgilmore, probably scop -- he knows a lot of the
push-script details 
(10:17:39) c4chris: Is there a plan to remove the group tag in spec
files?
(10:17:56) dgilmore: c4chris: not that i know of
(10:18:02) abadger1999: c4chris: Packaging Committee talked about it
today.
(10:18:03) tibbs: Yes.
(10:18:05) bpepple: c4chris: I think it's only be discussed so far.
(10:18:29) c4chris: So it's on the PC agenda?
(10:18:46) abadger1999: Won't happen for FC6 but we're coming up with a
timeline to make it optional and then dropped.
(10:19:03) c4chris: abadger1999, ok.  Great.
(10:19:35) c4chris: Another one is: can we have a hidden group in comps?
(10:20:01) c4chris: I think it'd be way easier to tell people: add all
your packages in comps
(10:20:19) thl: that was my stupid idea -- I think there were hidden
groups possible in the past
(10:20:22) c4chris: (until the day we have a shiny package database,
that is... :-) )
(10:21:00) abadger1999: Would be even better to have a hidden attribute.
(10:21:15) c4chris: I'm afraid there's no automated way to decide which
packages should appear in comps, no matter how hard we try
(10:21:54) abadger1999: So if Fedora policy is not to have commandline
tools generally, you can still put your application in and properly
categorized but flagged to be hidden.
(10:21:56) jima: all my packages are non-gui; only one (dnsmasq) seemed
common enough for me to bother putting it in comps.
(10:22:41) thl: jima, maybe a "command line tools" group would be a good
idea
(10:22:44) c4chris: IMHO, comps is a way to make some publicity for your
package
(10:22:56) rdieter: just put 'em all in there then (at least for now
until a better solution magically appears)
(10:23:02) jima: c4chris: if there's a category your package fits in.
(10:23:03) c4chris: you went to the trouble of packaging it: why not
advertise it somehow
(10:23:17) thl: c4chris, +1
(10:23:35) thl: c4chris, we IMHO really need a real comps SIG
(10:23:37) jima: (admittedly i only have one package that didn't fit
perfectly into a category)
(10:23:46) thl: or at least some people that really take care of it
(10:23:57) thl: there were so many questions on f-e-l in the past days
(10:24:02) c4chris: thl, yea probably
(10:24:02) abadger1999: c4chris: Is there a plan to have groups in the
package database?  (Or a desire?)
(10:24:09) thl: or should the PC handle comps?
(10:24:28) c4chris: thl, it affects core too...
(10:24:50) thl: yes and no
(10:24:53) c4chris: abadger1999, not completely sure
(10:25:11) thl: we IMHO need agroup that handles the comps stuff for
extras and watches it over
(10:25:11) abadger1999: It's not listed on the pages yet -- os if it's a
desire, it needs to get listed.
(10:25:22) bpepple: thl: +1
(10:25:22) abadger1999: s/os/so/
(10:25:22) c4chris: but the package database would be a nice place to
put such attributes I think
(10:25:39) rdieter: SIG++, it needs/deserves extra tlc
(10:25:51) rdieter: it = comps
(10:25:58) c4chris: tlc?
(10:26:07) rdieter: tender loving care.
(10:26:14) c4chris: Oh :-)
(10:26:22) c4chris: yes!
(10:26:25) thl: c4chris, well, seems you have a new job
(10:26:39) thl: c4chris, can you find one or two more poeple for the
sig?
(10:26:46) c4chris: thl, k
(10:26:49) ***bpepple would be willing to help/
(10:26:53) c4chris: Yup, I'll need help
(10:26:59) c4chris: bpepple, thanks
(10:27:17) bpepple: no prob.
(10:27:17) thl: c4chris, anything else regarding comps we should
discuss?
(10:27:31) c4chris: thl, no I'm fine
(10:27:40) c4chris: I'll start a wiki page soon(ish)
(10:28:11) thl: k, so let's move on
(10:28:23) thl: c4chris, can you update the status page on the schedule
please?
(10:28:33) thl: to reflect the recent discusssions?
(10:28:35) c4chris: thl, will do
(10:29:00) thl has changed the topic to: FESCo meeting in progress --
Activate legacy in buildroots
(10:29:12) thl: dgilmore told me that it's not activated yet
(10:29:20) thl: so I think we can skip this today
(10:29:22) dgilmore: not yet
(10:30:01) dgilmore: hopefully this week we can get everything in place.
i need to setup legacy to use the buildsys first  or they wont have fc4
ppc updates
(10:30:13) thl has changed the topic to: FESCo meeting in progress --
CTRL-C problem
(10:30:17) thl: skipping as well
(10:30:20) thl: ohh
(10:30:25) thl: or, no
(10:30:38) thl: warren, did you bring it up in the last infrasctructure
meeting?
(10:30:51) warren: bad news
(10:31:13) warren: It is the judgement of the infrastructure team that
this simply is not a priority.
(10:31:26) warren: If someone has ideas they're willing to try it.
(10:31:35) warren: that is all.
(10:31:52) thl: warren, not nice, but life sucks sometimes
(10:32:02) thl: warren, related to this
(10:32:14) thl: how is the schedule for the new VCS?
(10:32:25) thl: are there detailsed plans yet?
(10:32:30) warren: Infrastructure rather focus efforts on making VCS
happen sooner than to improve the existing one.
(10:32:43) thl: otherwise it'll be FC( until it's in place...
(10:32:50) warren: late August two new servers will be installed, and we
will simultaneously test mercurial and bazaar-ng
(10:33:35) tibbs: That's great news.
(10:33:59) warren: we're ordering a pretty sweet beefy box today
(10:34:03) thl: warren, yeah, great
(10:34:09) dgilmore: thl: from memory  we hopped to have something ready
for testing when FC7 test1 hits the streets
(10:34:11) abadger1999: I've got the backend for a bzr implementation
but we also need work done on the packaging database to work on a
finished front end.
(10:34:27) warren: 2x2 xeon, 8GB RAM, big SCSI drives.  It should host a
few xen guests comfortably for greater infrastructure flexibility.
(10:34:41) jima: nice!
(10:34:53) thl: k, let's stop here now and move on
(10:35:06) thl has changed the topic to: FESCo meeting in progress --
Packaging Committee Report
(10:35:34) abadger1999: We started discussion of removing the Group tag
today.
(10:35:54) abadger1999: Approved trying to get a patch into rpm to make
the tag optional for FC6.
(10:36:20) abadger1999: We'll look at the timeline for changing the
guidelines based on when that makes it into rpm.
(10:36:47) jima: if it's in by the time we're supposed to do the final
mass-rebuild, i wouldn't have any qualms with removing the tag from my
specs while i'm incrementing.
(10:37:03) abadger1999: ldconfig wording on the Guidelines page was
clarified but with the same meaning
(10:37:08) c4chris: sounds cool
(10:37:11) tibbs: jima: We're not targeting FC6 for this.
(10:37:46) abadger1999: That's all for changes.
(10:37:54) thl: abadger1999, thx
(10:38:01) abadger1999: Oh -- we're talking about changing meeting
date/time
(10:38:10) abadger1999: But we're continuing that on the mailing list.
(10:38:22) thl: I saw it roughly
(10:38:29) thl: there was the problem with DST
(10:38:50) thl: weco meets at 18:00 UTC during winters IIRC
(10:38:57) thl: and 17:00 during summers
(10:39:10) thl: so the effective meeting time stays the same
(10:39:28) thl: s/weco/FESCo/
(10:39:43) thl: is that scheme fine for all the new FESCo members, too?
(10:39:49) bpepple: Yup.
(10:39:56) abadger1999: Yes.
(10:40:00) c4chris: it's ok
(10:40:03) tibbs: Yes.
(10:40:14) rdieter: ok
(10:40:16) thl: k
(10:40:36) thl has changed the topic to: FESCo meeting in progress --
Sponsorship nominations
(10:40:41) thl: any new nominations?
(10:41:02) jwb: are we supposed to send them to FESCo list first?
(10:41:06) thl: btw, someone really should send Nodoid a summary why we
didn#t apporve him last time
(10:41:18) c4chris: I don't get mails from the sponsor list
(10:41:21) thl: does anyone still have the mails that were send around
when we discussed it?
(10:41:39) tibbs: I probably do.
(10:41:46) thl: I'd like to porpose that we send it to both lists
(10:41:56) thl: e.g. FESCo and sponsors
(10:42:05) bpepple: thl: +1
(10:42:07) c4chris: thl, yup, that's be useful
(10:42:08) thl: or we need to create a special mailinglist
(10:42:16) thl: but that's probaly overkill
(10:42:24) warren: any self-nominations?
(10:42:39) thl: I'm wonering if we should make c4chris a sponsor
(10:42:48) thl: FESCo members IMHO should also be sponsors
(10:42:51) tibbs: I have everything sent to
cvsextras-sponsors fedoraproject org since the beginning of May, BTW.
(10:43:25) thl: tibbs, can you send me the discussions around upgrading
nodoid? then I'l send him a summary why we didn't approve him
(10:43:35) tibbs: Going by the top reviewiers, Patrice Dumas would up
for sponsorship.
(10:44:36) dgilmore: thl: honestly  you could make me a sponser.  but i
dont know if i would be a good one
(10:44:41) thl: tibbs, can you send a mail to the list to start the
discussions
(10:44:51) dgilmore: tibbs: thats bad  not all of fesco get that
(10:45:11) thl: dgilmore, you don't have to use your powers if you don't
feel compfortable with it
(10:45:33) thl: who else from FESCo isn't a sponsor? Currently c4chris
and dgilmore afaics
(10:46:09) thl: a lot of silence here
(10:46:18) c4chris: I guess we are the only 2...
(10:46:24) thl: seems poeple don't like the iea to make all FESCo
members sponsors...
(10:46:28) dgilmore: thl: i guess its just me and c4chris 
(10:46:52) thl: s/iea/iea/
(10:46:53) c4chris: the thing is I don't do that many reviews
(10:46:56) thl: s/iea/idea/
(10:46:57) abadger1999: It's overloading the sponsorship role...
(10:47:13) jima: i don't see the justification, personally.
(10:47:27) thl: abadger1999, okay, so let's just drop that idea of mine
(10:47:33) jima: sponsors become sponsors based on merit, don't they?
(10:47:55) c4chris: yup, that's the idea
(10:48:03) thl: jima, yes, but that merit doesn't always mean "reviews"
(10:48:06) rdieter: Hopefully, FESCo implies merit... (:
(10:48:12) c4chris: and a deep knowledge of the packaging rules...
(10:48:13) thl: I didn't do to much reviews
(10:48:15) dgilmore: i guess being in Fesco  means your trusted by the
community
(10:48:20) thl: but I'm here and a sponsor, too
(10:48:23) BobJensen is now known as BobJensen-Away
(10:48:25) warren: I'm for giving them sponsorship, I trust that they
would do the right thing.
(10:48:34) rdieter: +1
(10:48:38) thl: +1
(10:48:41) bpepple: warren: +1
(10:48:54) abadger1999: I would give them sponsorship too -- but that's
separate from making all FESCo members sponsors.
(10:49:21) abadger1999: So c4chris, dgilmore sponsorship +1
(10:49:23) thl: abadger1999, yeah, your right
(10:49:25) thl: abadger1999, +1
(10:49:37) c4chris: So you need to simple rule to always nominate DESCo
members to sponsorship, and then simply debate it like usual...
(10:50:02) jima: c4chris: that sounds better to me than just
automatically upgrading them.
(10:50:03) c4chris: s/D/F/ doh
(10:50:15) tibbs: But the debate wouldn't necessarily be based on the
number and quality of reviews.
(10:50:20) jima: if they have merit, that's fine, imo.
(10:50:38) thl: I'm counting 5 votes to make c4chris and dgilmore
sponsors currently (four indirect)
(10:50:49) thl: so I think they are accepted
(10:50:58) thl: please yell now if you disagree
(10:51:02) jima: otherwise sponsorship can be attained by winning a
popularity contest (the fesco election) :)
(10:51:03) tibbs: I don't have any problem with having more sponsors,
but I am curious that we seem to have changed the qualifications.
(10:51:38) thl: tibbs, as I said -- I also didn't do to much reviews
(10:52:11) tibbs: Perhaps we can consider the qualifications at a future
meeting.  Because honestly we do need more active sponsors.
(10:52:27) thl: tibbs, yeah
(10:52:35) c4chris: tibbs, wouldn't be a bad idea
(10:52:47) tibbs: In any case, +1 for both folks.
(10:52:57) c4chris: that reminds me we need to discuss responsibilities
at some point...
(10:52:58) thl: I consider them accepted them
(10:53:11) thl: s/them/then/
(10:53:23) thl: there must be something wrong with my keyboard today ;-)
(10:53:23) tibbs: Do remember that you have to review in order to
actually sponsor someone.
(10:53:28) thl: so let's move on now
(10:53:50) thl has changed the topic to: FESCo meeting in progress --
approve kmod's
(10:54:01) thl: I didn#t get new request for approvals
(10:54:13) c4chris: no wonder...
(10:54:14) thl: and I didn#t find time to forward the zaptel issue to
FAB
(10:54:31) tibbs: Wasn't new kmod approval put on hold?
(10:54:41) thl: let's get the other stuff around kmods sorted out first
before we bring that on hte table again
(10:55:03) c4chris: yea, my inbox needs a break...
(10:55:12) bpepple: no doubt.
(10:55:14) thl: :)
(10:55:17) thl: so let's move on
(10:55:30) thl has changed the topic to: FESCo meeting in progress --
MISC from schedule
(10:55:41) thl has changed the topic to: FESCo meeting in progress --
MISC from schedule --  Proposal for (FC and) FE Package Review Request
template
(10:55:47) thl:
https://www.redhat.com/archives/fedora-extras-list/2006-August/msg00361.html
(10:55:50) thl: do we like the idea?
(10:56:04) c4chris: yes
(10:56:15) tibbs: Yes, the summary would be nice to have.
(10:56:22) abadger1999: Makes sense to me.
(10:56:27) tibbs: It would also be nice if people knew to remove the
angle brackets.
(10:56:44) c4chris: tibbs, yes
(10:56:45) ***bpepple doesn't have a problem with it.
(10:56:47) tibbs: I've found that it's not nearly as obvious to some
folks as you'd think.
(10:57:07) c4chris: maybe remove them from the template?
(10:57:27) thl: c4chris, +1
(10:57:47) thl: so we consider this accepted?
(10:57:49) bpepple: c4chris: +1
(10:57:57) c4chris: thl, +1
(10:58:03) rdieter: +1
(10:58:03) tibbs: +1
(10:58:06) bpepple: thl: +1
(10:58:14) thl: k, moving on
(10:58:21) thl has changed the topic to: FESCo meeting in progress --
MISC from schedule --  create a FESCO user in the wiki that sends mails
to the FESCo list and subscribe it to /Extras/Schedul.* in the wiki
(10:58:29) thl: I send this to the fesco list
(10:58:30) c4chris: who can update the template?
(10:58:41) bpepple: thl: +1, this sounds like a good idea.
(10:58:56) tibbs: Yes, this is definitely good.  +1
(10:59:06) thl: c4chris, djb (or what was his nick?); I'll take care of
it
(10:59:07) c4chris: why not.  I'm subscribed, but that would give us a
log of things
(10:59:17) c4chris: thl, k thx
(10:59:18) abadger1999: +1
(10:59:26) rdieter: sounds good to me, +1
(10:59:34) c4chris: (for the next time the wiki crashes... :-P)
(10:59:40) warren: Just do it 
(10:59:47) thl: k
(11:00:06) thl has changed the topic to: FESCo meeting in progress --
MISC from schedule --  jwb suggested in #fedora-extras that we discuss
the current kmod discusions
(11:00:06) warren: would that user's e-mail address be the fesco list ?
(11:00:16) thl: warren, yes
(11:00:32) thl: I'd like to stay mostly quiet on the kmod stuff
(11:00:34) thl: jwb, ?
(11:00:49) jwb: actually
(11:00:55) warren: The kmod issue.... is big and complicated.  For this
reason we scheduled tomorrow's meeting.  But even then I don't think we
will be successful.
(11:01:16) warren: I am uncertain what to do at this point.
(11:01:20) dgilmore: warren: ? toomorrows meeting
(11:01:32) jwb: warren, yeah... what dgilmore said
(11:01:47) tibbs: There's some sort of telephone conference scheduled.
(11:01:57) warren: thinking to do it on IRC instead
(11:01:57) jwb: with who?
(11:01:59) warren: haven't decided yet
(11:02:06) warren: see fedora-packaging for details
(11:02:18) jwb: kmod is bigger than just packaging
(11:02:25) warren: yes it is
(11:02:50) jwb: we've got davej and dwmw2 saying they should die in
Extras and Core all together
(11:03:00) jwb: among others
(11:03:02) dgilmore: I think that kmods are important for testing and
proving things. but that perhaps it should have its own repo
fedora-kmods  make the user have a clear decission to make knowing  that
core kernel developers wont support them
(11:03:28) jwb: dgilmore, that's what 3rd party repos are for
(11:03:49) jwb: dgilmore, IMHO, an "unsupported" repo by the kernel devs
cannot be part of the fedora project officially
(11:03:54) warren: Those kinds of decisions are FPB's to make.
(11:04:13) jwb: warren, that doesn't mean we don't get to make
suggestions
(11:04:41) abadger1999: If the guidelines are just for other repos, then
the whole landscape of what's a good guideline changes.
(11:04:55) jwb: agreed
(11:05:23) jwb: i think this is a case where we either care enough to
allow it in fedora repos, or we get off the pot and let 3rd party repos
decide what they want to do
(11:05:28) warren: If you truly care about the kmod issues, then please
seriously participate in the fedora-packaging list discussions.
(11:05:32) dgilmore: i think we could have a repo within fedora thats
disablled by default  where they can live
(11:05:39) warren: Right now it is a big mess.
(11:06:06) jwb: warren, i care more about having/not having modules in
fedora to begin with
(11:06:20) jwb: warren, i think that needs to be settled before any kind
of packaging standard comes about
(11:06:50) warren: OK, then that is defnitely something you need to
bring to FAB
(11:06:52) warren: are you on FAB?
(11:07:06) jwb: which i realize is confusing because kmods already
exist... where were davej and dwmw2 when kmods originated?
(11:07:10) jwb: warren, yes.  i can email FAB
(11:07:30) warren: jwb, package committee and fesco decides mainly on
technical details, you are asking about quasi-political policy
(11:07:36) thl: jwb, davej was asked by jeremy for permission of the
kmod stuff iirc
(11:07:44) jwb: thl, that's what i thought
(11:07:49) rdieter: gotta run, another meeting (fun).
(11:07:52) jwb: ok, i'd like to take a vote really quick
(11:07:59) rdieter: (ok, I'll wait)
(11:08:02) jwb: who in FESCo thinks we should have kmods?
(11:08:12) rdieter: have, vs. not have?
(11:08:17) dgilmore: jwb: my vote is for a kmod repository
(11:08:18) jwb: rdieter, right
(11:08:22) rdieter: have +1.
(11:08:27) tibbs: I'm undecided, sorry.
(11:08:50) thl: from the political standpoint: not have
(11:08:57) c4chris: I think we can have modules in Extras
(11:08:59) ***rdieter runs...
(11:09:04) rdieter is now known as rdieter_away
(11:09:10) skvidal: rdieter_away: come back here
(11:09:13) thl: but we compete with ubuntu and suse: and they have all
the stuff, so we should have it, too
(11:09:25) bpepple: have +1
(11:09:36) rdieter_away: skvidal: huh?
(11:09:41) warren: I think the current kmod standard with its strict
restrictions is a generally good thing.
(11:09:53) dgilmore: i think we can  but if its in its own repository
then users will be more aware of what support they can expect
(11:09:56) jwb: so in general, FESCo feels kmods should remain
(11:10:01) warren: It isn't a "free ride" into Fedora with any kmod.  it
must satisfy requirements, and pressure is put to push things upstream.
(11:10:02) thl: warren, maybe we shopuld put the "time restiction" back
on the table, too
(11:10:16) thl: e.g. allow each kmod in extras only for a certain time
(11:10:20) rdieter_away is now known as rdieter
(11:10:21) ***cweyl scrolls back and reads
(11:10:24) tibbs: I think that from a user's standpoint having access
under the Fedora umbrella to every module which does not violate the law
is a good thing.
(11:10:24) thl: three releases
(11:10:45) thl: tibbs, we should work towards modules that get merged
upstream
(11:10:59) dgilmore: warren: which can only be done in a fedora project
controlled repo.  3rd party repos == "no input from us"
(11:11:00) tibbs: But maintenance is paramount and the argument of the
kernel devs atainst having to deal with bug reports arising from
external modules is compelling.
(11:11:11) thl: It's really important that they get merged into the
vanilla kernel
(11:11:16) tibbs: thl: I disagree; I don't think it's our business to
push any code author towards anything they don't want to do.
(11:11:56) thl: tibbs, http://www.fedoraproject.org/wiki/Objectives : To
do as much of the development work as possible directly in the upstream
packages.
(11:12:12) bpepple: thl: +1
(11:12:16) tibbs: I don't see how that applies.
(11:12:28) tibbs: A kernel module's upstream isn't necessarily the
kernel.
(11:12:32) dgilmore: thl: yes  but a kmod upstream is different to
kernel upstream
(11:12:40) abadger1999: tibbs: +1
(11:12:48) tibbs: That objective just says we shouldn't carry around
fedora-specific patches in fedora packages.
(11:13:06) cweyl: tibbs: +1 on "...is a good thing"
(11:13:08) dgilmore: thl: if we patch a kmod  we should get it upstream
whereever that is
(11:13:16) warren: The way spot describes it, the restrictions set by
packaging committee on what qualifies for kmod is strict.
(11:13:34) jwb: thl, i'll email FAB asking for a political decision
(11:13:42) thl: jwb, thx
(11:13:49) ***jwb steps away for a second
(11:13:51) thl: so let's stop here for today
(11:14:02) warren: The time requirement is not exactly important to
decide now, it is closer to a political decision.  The bigger problem is
the technical issues for the near-term.
(11:14:03) thl has changed the topic to: FESCo meeting in progress --
free discussion
(11:14:16) tibbs: I started on
http://fedoraproject.org/wiki/JasonTibbitts/MaintainerResponsibilityPolicy
(11:14:16) thl: anything else we should discuss?
(11:14:40) tibbs: A nice rat's nest there which will require a good bit
of spirited discussion.
(11:14:56) tibbs: I'd like to see anyone interested contribute to that
document.
(11:15:34) thl: tibbs, can you move that over to the FESCo namespace
please
(11:15:36) tibbs: Unfortunately my wiki-fu is still crap and I can't get
the list indentation right.
(11:15:44) thl: then I'll create a entry on the schedule
(11:15:50) rdieter is now known as rdieter_away
(11:16:00) tibbs: Where would you like it to live?
(11:16:08) tibbs: Under Schedule?
(11:16:45) thl: tibbs, let me handle the moving
(11:16:55) thl: I can also look at the list indentation
(11:17:16) cweyl: tibbs: I assume rabble contributions are good as well?
(11:17:24) tibbs: I was trying to get too fancy with boxed bits within a
list.
(11:17:36) tibbs: cweyl: I want to see as much input on this kind of
thing as is possible.
(11:17:42) cweyl: cool.
(11:17:42) c4chris: cweyl, sure
(11:18:13) cweyl: not that I've ever asked before.... ;)
(11:18:29) c4chris: FYI I won't be here next week (vacations, yay! :-) )
(11:18:37) jima: err, can rabble edit things under FESCo namespace?
(11:18:47) thl: jima, yes
(11:19:02) jima: ok, wasn't sure if it was locked down like Packaging
(11:19:09) cweyl: yah.  same here
(11:19:20) tibbs: Packaging is an anomaly, I think.
(11:19:32) jima: ok
(11:19:40) tibbs: You just need to be in EditGroup.
(11:20:25) thl: k, anything else?
(11:20:31) ***thl needs to leave soon
(11:20:41) thl: abadger1999, btw, many thx for writing the summaries
(11:20:48) tibbs: Nothing else from me.
(11:20:53) ***thl will close the meeting in 60
(11:21:17) abadger1999: jima: BTW: PackagingDrafts shouldn't be locked
down, just the actual Packaging hierarchy.
(11:21:21) ***thl will close the meeting in 30
(11:21:28) abadger1999: thl: np.
(11:21:41) ***thl will close the meeting in 10
(11:21:53) thl: -- MARK -- Meeting end
(11:21:57) thl: thx everyone!
(11:22:03) c4chris: thl, thx.
}}}

Attachment: signature.asc
Description: This is a digitally signed message part


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]