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

2006 August 05 FESCo Meeting Minutes

= 2006 August 05 FESCo Meeting =

Meeting Summaries are posted on the wiki at:


== Attending ==
 * thl
 * c4chris
 * tibbs
 * bpepple
 * abadger1999
 * warren
 * jwb (late)
 * dgilmore (late)

== Summary ==
 * Mass Rebuild
  * Builders are updated.  Need to add python and elfutils so that FC <
5 works in the builders.
  * Mass rebuild for FC6 will start on Mon, 28th August since Core is
not expecting any changes that would require a rebuild after that.
  * Email to fedora-extras will request comments on length of time for
maintainers to rebuild their packages before someone else rebuilds for
them and whether all packages need to be rebuilt or not.
 * Comaintainership
  * thl sent a message but got no replies so far.
  * Having SIGs be co-maintainers was well received.
  * Package database is needed to enable a lot of the functionality.
  * SCM needs (should?) have ACLs for separating write access by group.
This is hopefully coming between FC6 and FC7.
 * Comps.xml
  * For now the report on comps will list everything except perl-*,
python-*, lib*, and  *-devel.
  * Better results but harder to implement would be to retrieve
dependency information and only complain about packages which are not
depended upon by anything else.
 * Stalled Review
  * Approved:
  * Warren will setup nobody fedoraproject org to handle unassigned
 * Activate Legacy in the buildroots:
  * Most FESCo members accepted that legacy should be added to the
buildroots (has security fixes and is what we want end-users on Legacy
to run.)
  * There was some discussion around mschwendt's post about this:

  * Dissent centered around who is responsible for supporting FE when
the equivalent FC goes Legacy.
   * Security Team, Fedora Legacy (Core), and at least some Fedora
Extras maintainers don't want the responsiblity.
   * Co-maintainers (with the package database) could help track
volunteers who are fixing the older releases but this still leaves open
who is responsible.
   * At the moment, some FE maintainers are doing the work but dgilmore
has personally done ~10 security rebuilds on FE3 for packages he does
not own.
  * tibbs will bring this to extras-list as a draft document of
maintainers responsibilities.
 * Kmod reviews
  * sysprof was accepted as it seems to be an interim measure until the
userspace GUI can be migrated to oprofile.
   * However, there might not be any work being done on this currently.
  * zaptel: The upstream authors have stated they do not want to get the
module into the kernel due to their preference for a dual license.
   * nirik will generate an email about zaptel in specific and kmod
criteria in general on the fedora-devel list.
 * CVS Branch Requests
  * As an interim measure, packagers submitting a branch request should
append the bug number to the request:
      {{{* FC-4 FC-5 cluestick 28437}}}
  * Package Database is the eventual solution.
 * Package Database
  * c4chris will open a discussion on fedora-extras this week.
 * Encourage Extras Reviews
  * We're doing better now.
  * Getting new people involved via co-maintainership should help as
  * This item is being closed for now.
 * mono packaging
  * mono-core may move to libdir in time for FC6 -- it depends on how
serious this bug turns out to be:

== Log ==

(10:00:23) thl: my clock says it's time to start
(10:00:35) c4chris: mine too :)
(10:00:37) thl has changed the topic to: FESCo Meeting in progress
(10:00:48) thl: k, who's around?
(10:00:57) tibbs: I'm here.
(10:00:58) ***bpepple is here.
(10:00:59) ***cweyl is here (rabble)
(10:01:03) abadger1999: i'm back
(10:01:24) ***nirik is in the rabble seats. 
(10:01:35) thl: dgilmore, are you around?
(10:02:01) thl: well, let's start slowly
(10:02:05) c4chris: hmm 5/13 th...
(10:02:43) warren: here
(10:02:43) thl: c4chris|w, well, I prefer a meeting with only 5 people
over no meeting
(10:03:08) c4chris: sure.  should've added a half smiley
(10:03:11) thl: c4chris, we did a lot of meeting with only five of 17 in
the old FESCo
(10:03:18) thl: c4chris, k
(10:03:34) c4chris: 6 now :-)
(10:03:40) thl: I got the impression that some people only want meetings
when we are at least seven ;-)
(10:03:43) thl has changed the topic to: FESCo Meeting in progress --
(10:03:46) ***nirik would hope that the shiny new FESCo would have more
folks interested in participating....
(10:04:01) tibbs: Several folks are on vacation this week.
(10:04:09) tibbs: At leasr Rex and spot.
(10:04:10) thl: k, dglimore upgraded the builders
(10:04:21) thl: that's not working perfectly
(10:04:24) warren: Is test3 enough time to rebuild everything before
(10:04:24) thl: but mostly
(10:04:46) thl: warren, I still prefer to start a bit earlier
(10:04:46) c4chris: thl, anyone working on the strip problem?
(10:04:57) thl: c4chris, I suppose we need dgilmore for that job
(10:04:57) tibbs: Is there any downside to having folks start early if
they want?
(10:05:00) warren: c4chris, is that documented anywhere?
(10:05:14) abadger1999: tibbs: Just that they might have to do it again
(10:05:17) thl: warren, on the list only currently
(10:05:20) c4chris: warren, no idea
(10:05:30) abadger1999: (If there's a late gcc change, python upgrade,
et al.)
(10:05:38) thl: abadger1999, agreed
(10:06:00) thl: warren, "late gcc change, python upgrade, et al." -> if
we get a go from core we can start
(10:06:13) warren: let me check
(10:06:41) stickster [n=paul] entered the room.
(10:07:07) thl: warren, but I prefer FC6T3 over a "ohh, sorry, we found
a problem in foo; you need to rebuild everything in Extras once again"
(10:07:10) c4chris: dgilmore sent a messaeg saying he'd be at ClueCon
(whatever that is)
(10:07:24) bpepple: thl: +1
(10:07:38) nirik: FYI, the stripping and python issues for fc4 are
discussed in this thread:
(10:08:14) warren: No changes are expected, we should go ahead.  How
about late August?
(10:08:28) thl: warren, sounds good to me
(10:08:42) thl: 28 August?
(10:08:49) thl: that's a monday
(10:09:02) cweyl: is there a definition of what needs to be rebuilt,
again?  all packages !noarch?
(10:09:23) thl: I don't have a definition yet
(10:09:32) thl: scop probably should know
(10:09:48) warren: how about we discuss that on list, and have a
definition by August 24th's meeting
(10:10:08) thl: warren, +1
(10:10:14) thl: but one thing here
(10:10:30) thl: how much time should we give people to rebuild there
(10:10:33) thl: two weeks?
(10:10:49) thl: three weeks?
(10:10:51) bpepple: thl: How about 3?  Some people have got a lot of
(10:10:55) c4chris: I think 3
(10:11:01) tibbs: It kind of depends on how the buildsys can handle the
(10:11:13) tibbs: The new buildroots seem to have sped things up a bit,
which is good.
(10:11:29) tibbs: But if we experience the hanging problems again then
it could take quite some time.
(10:11:37) c4chris: Did we ever bring the buildsys down through
(10:12:07) thl: c4chris, I don#t think so
(10:12:22) abadger1999: I don't think so but there were deadlocks in the
code previously.
(10:12:27) cweyl: jobs should just queue up once the buildsys has
reached it's limit, right?
(10:12:30) stickster left the room ("Withdrawing to the lair...").
(10:12:31) abadger1999: with those gone, we can find new bugs :-)
(10:12:40) ***cweyl thinks this is going to be an interesting stress
test of plague :)
(10:12:40) warren: deadlocks were signal based at first, then kernel
(10:12:53) c4chris: sure, the code had problems, but I think the system
handles the load nicely
(10:13:31) thl: okay
(10:13:43) thl: let's think about it and discuss the other details in
later meetins
(10:13:50) thl: moving on
(10:14:01) tibbs: I'm not disparaging the buildsys, just noting that if
there are problems we should consider plan b.
(10:14:02) c4chris: who open the thread on fel?
(10:14:08) thl has changed the topic to: FESCo Meeting in progress --
(10:14:22) thl: c4chris, seems scop was interesed ;-)
(10:14:36) c4chris: scop +1 :-)
(10:14:41) tibbs: scop had to be away today....
(10:14:46) thl: well, I got no replys for my Comaintainership mail
(10:14:53) c4chris: tibbs, I know... ;-)
(10:15:02) ***bpepple meant to write something, but hasn't found time.
(10:15:08) thl: so what do to now? ifnore it?
(10:15:10) thl: ignore
(10:15:14) cweyl: thl: I liked the idea of SIGs having comaintainership
over their packages
(10:15:23) c4chris: I mostly liked it.
(10:15:25) tibbs: BTW, thl, the new schedule looks great.
(10:15:42) c4chris: I think we need the package database to make things
really work.
(10:15:43) cweyl: it makes sense to me, they'd likely have both a vested
interest in the packages and the knowledge of how to do it right
(10:15:46) thl: tibbs, thx; I also hope it's a bit easier to keep it
up2date and informative this way
(10:15:59) cweyl: c4chris: +1 for package database
(10:16:01) thl: cweyl, yes, that's one of the things we need
(10:16:12) thl: s/cweyl/c4chris/
(10:16:16) warren: Technically this is no different from who was allowed
to make changes in the past.  Only difference is creating formalized
groups.  Similar vein to multiple owners.
(10:16:41) warren: doesn't this count as another aspect where the
package database would be handy?
(10:17:10) c4chris: and apparently we'd need some support from the SCM
(10:17:23) c4chris: warren, yes I think so
(10:17:25) thl: warren, yes, it would be very handy
(10:17:37) thl: but we need support from the SCM, too
(10:17:49) warren: how so?
(10:18:00) abadger1999: All you need is ACLs for groups, yes?
(10:18:02) c4chris: to manage who can change the repo
(10:18:20) warren: We generally plan on keeping checkin access almost
wide open
(10:18:26) thl: warren, read
(10:18:26) warren: ACL groups will happen though, yes.
(10:18:53) thl: ACL will happen in CVS or the new SCM?
(10:18:59) giallu left the room (quit: Read error: 110 (Connection timed
(10:19:10) warren: Ideally new SCM
(10:19:14) thl: abadger1999, yes, I think all we need are  ACLs for
groups, yes
(10:19:22) c4chris: ETA ?
(10:19:35) thl: warren, yeah, ETA?
(10:19:44) warren: The entire new SCM thing is post-FC6
(10:19:46) abadger1999: warren: Could I get someplace to put my bzr
proof of concept?
(10:20:05) warren: abadger1999, good question... we should discuss that
during infrastructure meeting
(10:20:11) c4chris: warren, and pre FC& ?
(10:20:22) abadger1999: warren: I'll add it to the schedule there.
(10:20:23) warren: c4chris, that is the hope.
(10:20:26) c4chris: s/FC&/FC7/
(10:20:49) c4chris: k
(10:20:55) thl: let's stop here
(10:21:07) thl: maybe I get some more replys to the list for my mail ;-)
(10:21:08) warren: I think people have the wrong idea here.
(10:21:16) thl: warren, ?
(10:21:23) thl: warren, did you actually read my mail?
(10:21:32) warren: The important part of this is multiple owners, groups
of owners, and subscription mapping between owners/groups and packages.
(10:22:13) c4chris: warren, sure.  But what is the wrong idea ?
(10:22:50) warren: I guess it is the same idea, but different part is
(10:23:23) thl: warren, let's stop here and move on
(10:23:26) warren: ok
(10:23:33) thl: warren, maybe you can post your ideas to the list
(10:23:46) thl has changed the topic to: FESCo Meeting in progress --
Use comps.xml properly
(10:23:52) thl: c4chris ?
(10:24:10) thl: "c4chris evaluating what packages need to be listed"
(10:24:27) drfickle left the room (quit: "Please do not reply to this
(10:24:59) thl: c4chris lost :|
(10:25:02) c4chris: My current thinking is to get most packages except
perl- python- and -devel stuff
(10:25:13) c4chris: (sorry, slow typing here)
(10:25:29) c4chris: not sure about lib* stuff
(10:25:38) c4chris: probably not
(10:25:48) thl: c4chris, yeah, probably not
(10:25:55) c4chris: Of course, people are free to add those they like
(10:26:07) thl: c4chris, can you automate checks in you PackageStatus
(10:26:09) c4chris: I'll just not complain if they are not listed
(10:26:21) f13: if a package is a dep of something already in comps,
don't list it
(10:26:23) c4chris: thl, tes, that's the idea
(10:26:37) tibbs: Well, not necessarily.
(10:26:40) ***jwb is here now
(10:26:41) c4chris: f13, yes, that's a good idea
(10:27:07) c4chris: tibbs, sure, but don't have the script complain in
this case
(10:27:25) tibbs: OK, I see what you're saying.
(10:27:55) thl: c4chris, can you work further on this? there is no real
need to hurry
(10:28:05) thl: c4chris, but we should have something in place before
(10:28:13) c4chris: getting the dep tree from teh script might be a bit
tricky though...
(10:28:13) thl: e.g. mid-september?
(10:28:38) c4chris: thl, yes, I plan to get things into shape over the
next couple weeks.
(10:28:45) thl: c4chris, k
(10:29:00) thl: c4chris, is it ok for you if I move on? or do you want
anything discussed?
(10:29:28) c4chris: thl, fine with me.  Please post ideas to the thread
(10:29:36) thl has changed the topic to: FESCo Meeting in progress --
Stalled Package Reviews
(10:29:39) thl: thx c4chris 
(10:29:39) osphy [n=randey] entered the room.
(10:29:45) tibbs: Current draft is in:
(10:30:12) thl: tibbs, okay for me
(10:30:17) tibbs: I added the bit about FE-DEADREVIEW as suggested on
extras-list but otherwise haven't revised it since my initial post.
(10:30:18) thl: other opinions?
(10:30:39) c4chris: I like it
(10:30:39) bpepple: Looks alright to me.
(10:30:57) tibbs: No complaints from me.
(10:30:58) cweyl: I like it.  especially the FE-DEADREVIEW part ;)
(10:31:01) tibbs: (obviously)
(10:31:25) tibbs: I can set up the FE-DEADREVIEW blocker.
(10:31:29) thl: do we need to play the "+1" game or was that "I like it"
enough for everyone?
(10:31:49) tibbs: Did nobody fedoraproject org ever get set up?
(10:31:50) bpepple: Should be fine.  I didn't hear any complaints.
(10:31:50) jwb: i like it
(10:32:03) c4chris: no disagreements. All clear
(10:32:03) thl: warren, nobody fedoraproject org ?
(10:32:08) abadger1999: FE-DEADREVIEW was a nice addition.  I don't see
anything wrong with it.
(10:32:21) thl: okay, "Stalled Package Reviews" accepted
(10:32:23) warren: thl, oops, I forgot to follow through, I can do that
(10:32:31) mharris_sun [n=mharris] entered the room.
(10:32:34) thl: tibbs, can you integrate it into the wiki at the proper
place? tia!
(10:32:34) tibbs: I just need to know who to assign unassugned bugs back
(10:32:38) warren: mharris_sun, sun?
(10:32:49) tibbs: Since we can't put things back into NEW and there's no
(10:32:51) mharris_sun is now known as mharris
(10:32:56) thl: warren, no need to hurry
(10:32:59) mharris: warren: better? ;)
(10:33:14) thl: k, moving on
(10:33:24) thl has changed the topic to: FESCo Meeting in progress --
Activate legacy in buildroots
(10:33:25) tibbs: I'll just use the current phantom thl mailbox until
nobody fedoraproject org gets set up.
(10:33:37) thl: tibbs, I can create the bug, too
(10:33:51) thl: tibbs, otherwise you'll get all the mails also because
you are the reporter
(10:34:17) tibbs: I have my own phantom mailbox, so it's no problem, but
go ahead and create it if you like.
(10:34:43) dgilmore: hey guys  im here now
(10:34:45) thl: tibbs, okay; remond me if I forget about it 
(10:34:59) thl: ohh, welcome dgilmore :)
(10:35:06) xris [n=xris] entered the room.
(10:35:34) thl: dgilmore, did you read
https://www.redhat.com/archives/fedora-extras-list/2006-August/msg00091.html yet?
(10:35:43) dgilmore: thl: no
(10:35:49) dgilmore: let me read now
(10:35:55) thl has changed the topic to: FESCo Meeting in progress --
problems with the new builders
(10:36:17) dgilmore: thl: eithe elfutils is broken or not installed
(10:36:30) thl: dgilmore, read the thread completely
(10:36:31) dgilmore: so we need to install elfutils  if it is not
(10:36:40) thl: there is also a bug filed somewhere
(10:36:47) thl: we should try to fix that soon
(10:36:54) thl: dgilmore, are you still traveling?
(10:37:12) dgilmore: thl: im eating lunch at the conference
(10:37:27) dgilmore: network conectivity has been boorked
(10:37:28) abadger1999:
(10:37:50) abadger1999: The bug.  Details other problems with pre-FC5
buildroots as well.
(10:38:06) thl: dgilmore, when do you get home again?
(10:38:29) dgilmore: thl: tonight
(10:38:51) thl: dgilmore, can you fix the debug-problem this weekend
(10:39:09) tibbs: Does anyone know how the dependency coverage of the
minimal buildroot was calculated?
(10:39:34) warren: installing a chroot with yum
(10:39:49) warren: If things are missing, just add them to the list and
not worry about it.
(10:39:50) tibbs: OK, that makes sense.
(10:39:59) thl: tibbs, python probably should be added for FC3 and FC4
(10:40:34) tibbs: I kind of wish it were the other way, with python not
being in FC5+
(10:40:55) thl: tibbs, I think there is a pyhton dep in rpm (or
somewhere else) that pulls it in 
(10:40:57) tibbs: But with cached buildroots in mock now it's much less
of an issue.
(10:41:17) dgilmore: thl: i dont know if i can.  it will need to be
fixed in the buildsys-build package i dont know where it lives and where
i need to put it so we access it
(10:41:40) thl: dgilmore, skvidal should know
(10:42:18) thl: I'm getting more and more annoyed of this
"buildsys-build is somewhere in the web" scheme
(10:42:32) thl: I'd prefer if it would be shipped together with mock :-/
(10:42:43) cweyl: thl: +1
(10:42:58) dgilmore: thl: python  can easily be added  if we add the
macro for scop  and require fedora-rpmdevtools
(10:43:21) dgilmore: tibbs: we are not using cached buildroots
(10:43:47) c4chris: dgilmore, I thought we would...
(10:43:52) thl: hmmmm, so how to proceed
(10:44:09) tibbs: I am at home, which is where it counts for me since I
do so many builds.
(10:44:11) thl: dgilmore, let's us two try to ping skvidal tomorrow to
get this fixed
(10:44:19) dgilmore: maybe we could  and rebuild the cache with each
core update
(10:44:25) dgilmore: thl: sure
(10:44:31) thl: everything okay if we add elf-utils and python to the
minimal buildroots?
(10:44:49) thl: everybody
(10:44:57) dgilmore: im fine with it
(10:44:58) warren: It doesn't hurt to add redundant things to the list.
(10:45:01) c4chris: thl, +1
(10:45:01) tibbs: Can we just do it for the releases that need it?
(10:45:02) warren: Just Do it
(10:45:11) thl: tibbs, agreed
(10:45:28) dgilmore: tibbs: yes   we just need to maintain each distro
(10:45:37) c4chris: tibbs, k
(10:45:37) tibbs: Otherwise two years from now we'll argue about why it
was put there.
(10:45:46) tibbs: Just like most of the stuff we cut out last time.
(10:45:49) thl: okay, then this seems to be settled for now ; let's move
(10:45:56) thl has changed the topic to: FESCo Meeting in progress --
Activate legacy in buildroots
(10:46:00) thl: back to this one
(10:46:17) thl: mschwendt didn't like this idea
(10:46:22) thl: he posted to the list once
(10:46:23) c4chris: what needs to happen?
(10:46:42) tibbs: Legacy is dropping FC2- soon anyway; so don't we
pretty much have this already?
(10:46:43) thl: c4chris, well, we just need to say "We want to activate
legacy in buildroots"
(10:46:52) dgilmore: c4chris:make sure fc3 packages get built with
legacy updates  in the buildroot also
(10:47:12) c4chris: thl, meaning we pull packages from legacy into the
(10:47:18) thl: c4chris, yes
(10:47:39) dgilmore: mschwendt  mostly didint like it because if a user
does not have legacy installed  then  it might break something
(10:47:39) c4chris: k
(10:47:52) thl: here's the mail from mschwendt:
(10:47:59) thl: he has some points 
(10:48:06) dgilmore: I say  if your still running FC3 like i am,  then
your are an idiot if you are not using legacy updates
(10:48:23) thl: but I don#t think we have a choice
(10:48:35) thl: we IMHOI must activate legacy in the builders for FE3
(10:48:40) thl: and FE4 soon
(10:48:55) dgilmore: thl: I fully agree which is why i brought it up
(10:48:56) thl: other opinions?
(10:49:14) thl: or let's play the "+1" game again:
(10:49:22) tibbs: I agree that this needs to be done.  (Says someone
with loads of FC3 still around.)
(10:49:23) thl: +1 for activate legacy in buildroots
(10:49:31) c4chris: sounds saner to me
(10:49:33) dgilmore: +1
(10:49:35) c4chris: +1
(10:49:46) tibbs: +1
(10:50:32) warren: Just do it!
(10:50:32) abadger1999: -1
(10:50:34) ***thl will wait another 15 seconds before he considers this
(10:50:37) thl: ohh
(10:50:42) thl: abadger1999, --verbose please
(10:50:44) abadger1999: Did anyone answer mschwendt's points?
(10:51:01) abadger1999: Just to say -- here's why I think you're barking
up the wrong tree?
(10:51:05) thl: abadger1999, I started once, but got distracted
(10:51:16) dgilmore: thl: as did i
(10:51:40) c4chris: do we really want FE + FC versus FE + FL + FC ?
(10:52:02) abadger1999: heh -- So itsounds like you have answers to his
points but no one has put them in a public place where everyone can then
understand the reasoning.
(10:52:08) c4chris: there's supposed to be security fixes in FL
(10:52:20) thl: c4chris, "do we really want to build FE for a FC4 that's
unsupported without FL" 
(10:52:23) thl: ?
(10:52:35) c4chris: thl, yes
(10:52:50) tibbs: The breakage goes the other way, too.
(10:52:50) cweyl: c4chris: is it really "FC + FC + FL"?  I mean, e.g.,
FC-3 was "transferred" to FL...
(10:53:14) tibbs: do we really want to build FE for an FC4 that's
unsupported with FL?
(10:53:17) c4chris: cweyl, F[CL] probably
(10:53:25) thl: abadger1999, we probably should answer
(10:53:32) thl: abadger1999, but there are so many mails on the list
(10:53:47) thl: abadger1999, we can reply to each and everyone because
we never get stuff done otherwise
(10:54:04) cweyl: yah.  I mean, if FL is responsible for updates to
FC-3, then it seems like if we don't include them it's almost the same
as just disabling the updates repo
(10:54:37) dgilmore: cweyl: thats what i think
(10:54:42) c4chris: cweyl, that's my impression too
(10:55:10) thl: yeah, mine too
(10:55:13) abadger1999: thl: I think the one point from his emails I
want answered is who maintains FE-legacy?
(10:55:24) bpepple: abadger1999: +1
(10:55:25) thl: abadger1999, there is no FE-legacy
(10:55:29) abadger1999: Exactly.
(10:55:42) thl: abadger1999, when we created the EOL policy
(10:55:52) thl: most people wanted to avoid FE-legacy
(10:56:05) dgilmore: abadger1999: we maintaine FE in legacy
(10:56:06) thl: so the normal maintainers are responsible for their
(10:56:09) abadger1999: So if FC-Legacy upgrades a library to fix a
security hole and that breaks FE apps, who is going to fix it?
(10:56:28) dgilmore: abadger1999: the maintainer
(10:56:32) thl: abadger1999, the Extras maintainer should fix it
(10:56:34) tibbs: This came up recently with some security fixes.  One
maintainer wasn't sure if he was responsible for fixing the FE3 version
of his package.
(10:56:37) dgilmore: if not them  then the Fedora Security team
(10:56:40) abadger1999: No.
(10:57:06) tibbs: I told him that it's much preferred that he do it, but
that if he insisted on not doing it then the security team would step
(10:57:07) abadger1999: Maintainers have continuously said that they
don't want indefinite responsibility for their packages.
(10:57:12) dgilmore: abadger1999: yes  the maintainer  should.  he
should be strongly encouraged to do so
(10:57:32) dgilmore: abadger1999: if he doesnt  then i will do so
(10:57:50) thl: abadger1999, this could be solved with a package
(10:57:55) dgilmore: I have already build 10 or so other peoples fe3
packages for security reasons
(10:58:13) thl: abadger1999, then maintainer could hand over their
packages for older dists to other people
(10:58:14) cweyl: abadger1999: agreed.  there needs to be a point at
which a maintainer can legitimately say "that stuff is just too old.
you now get to keep both parts"
(10:58:28) thl: abadger1999, but I don#t cosider this a prople specific
to FL
(10:58:41) thl: abadger1999, I suppose there are already some
maintainers that ignore FE4 also
(10:58:54) thl: that's a general problem that needs to be solved
(10:58:55) abadger1999: thl: Not really (package db).  It doesn't assign
responsibility, it only records who has volunteered.
(10:59:22) abadger1999: thl: +1 to general problem.
(10:59:30) tibbs: Do we have a document on maintainer responsibility?
(10:59:31) dgilmore: abadger1999: there is no package DB   but it needs
to be one of the design goals  of the package db
(10:59:31) thl: well, weÄre running late
(10:59:34) cweyl: maybe something like "consider all extras branches for
legacy distros orphaned unless explicitly owned?  (e.g. via a
'maintained' file containing the email of the owner in the branch?)"
(10:59:39) bpepple: tibbs: Not that I'm aware of.
(10:59:43) ***cweyl is just brainstorming
(10:59:47) thl: someone really need to bring this to the list again for
further discussion
(10:59:50) tibbs: If not, I'm happy to start one.
(10:59:53) thl: any volunteers?
(10:59:58) tibbs: Lots of argument potential there.
(11:00:16) thl: tibbs, no, we have no defined "maintainer [19:59:29] <
tibbs> | Do we have a document on maintainer responsibility?
(11:00:26) thl: tibbs, no, we have no defined responsibilitys"
(11:00:42) thl: sorry, cut'n'paste by touchpad error
(11:00:58) tibbs: Is there agreement that we need to define some basic
(11:01:00) thl: tibbs, but we need one soon (er or later)
(11:01:04) thl: tibbs, yes
(11:01:06) bpepple: tibbs: +1
(11:01:20) c4chris: tibbs, +1
(11:01:29) abadger1999: tibbs: +1
(11:01:46) tibbs: I'll start writing something up and will take it to
the list for further discussion.
(11:01:53) thl: tibbs, tia
(11:01:58) tibbs: extras-list, right?  Not fesco-list?
(11:02:02) thl: tibbs, yes
(11:02:11) thl: extras-list
(11:02:18) tibbs: OK, I'll try to have something soon.
(11:02:32) thl has changed the topic to: FESCo Meeting in progress --
CTRL-C problem
(11:02:36) thl: no news yet afaik
(11:02:39) thl: skipping
(11:02:47) thl has changed the topic to: FESCo Meeting in progress --
IPv6 Support in Extras
(11:02:52) thl: jwb, ?
(11:03:00) thl: do we need to discuss this today?
(11:03:00) jwb: i think it has mostly been covered
(11:03:05) thl: It's late already
(11:03:10) thl: jwb, mostly?
(11:03:20) jwb: from a packaging standpoint, yes
(11:03:27) jwb: we talked about it a few weeks ago
(11:03:38) jwb: it's fine if we skip it today
(11:03:46) thl: jwb, the packaging comittee looked into this, too
(11:03:57) jwb: right, that's the discussion i'm referring to
(11:04:02) thl: jwb, they created some guidelines for stuff like this
(11:04:20) jwb: yep
(11:04:22) tibbs: Yes, we did, but there's still some question as to
whether it's really our business to define something like that.
(11:05:02) tibbs: Besides, the guidelines the packaging committee
adopted is quite weak.
(11:05:27) thl: well, let's skip it for today then
(11:05:39) thl has changed the topic to: FESCo Meeting in progress --
Packaging Committee Report
(11:05:41) thl: tibbs, ?
(11:05:56) abadger1999: Nothing to report from the Packaging committee.
(11:05:58) tibbs: Not a lot of folks around; nothing was able to be
voted upon.
(11:06:02) abadger1999: Not enough members present today.
(11:06:06) thl: okay, moving on
(11:06:15) thl has changed the topic to: FESCo Meeting in progress --
Weekly sponsorship nomination
(11:06:19) thl: any new nominations?
(11:06:40) ***thl will wait for about 30 seconds
(11:06:40) ***noddy pipes up that he would like to be considered again
for sponsor
(11:06:40) warren: brb, bathroom
(11:06:57) ixs: mhm
(11:07:01) ixs: are new sponsors needed?
(11:07:14) ***bpepple isn't sure.
(11:07:27) cweyl: ixs: they're always needed :)
(11:07:41) ixs: cweyl: ;D
(11:08:00) ixs: ahh well, I'll go next week. ;D
(11:08:14) tibbs: I wonder if c4chris couldn't be persuaded to generate
some stats on FE-NEEDSPONSOR.
(11:08:40) thl: noddy, we should probably wait some weeks before we
discuss upgrading you to a sponsor again
(11:08:43) thl: noddy, that okay?
(11:08:46) c4chris: tibbs, there be some.  What would you like to see?
(11:08:58) tibbs: I did a lot of sponsorship work last week and it
turned out rather well, I think.
(11:09:22) tibbs: c4chris: Just a count over time, I guess.  If we're
trending upwards then we need more sponsors or more sponsorship
(11:09:31) noddy: thl : not a problem. I'll call back in the middle of
Aug if you think that's long enough
(11:09:47) warren: back
(11:10:03) c4chris: tibbs, ok, I'll give this a shot
(11:10:14) thl: k, so let's move on
(11:10:18) tibbs: c4chris: I do see the NEEDSPONSOR section in
PackageStatus now; thanks.
(11:10:25) thl has changed the topic to: FESCo Meeting in progress --
review kmod's
(11:10:32) thl: the easy one first
(11:10:55) thl: sysprof
(11:11:01) thl: +0,75 from me
(11:11:12) c4chris: +1
(11:11:13) thl: it's not perfect
(11:11:19) thl: but acceptable IMHO
(11:11:47) thl: that all?
(11:11:54) tibbs: +1 as well; I still wish someone would make the
userspace part work with oprofile.
(11:12:02) tibbs: but I don't think that should keep it out.
(11:12:06) bpepple: +1
(11:12:09) abadger1999: +1
(11:12:18) tibbs: Note that scop voted +1 on the list.
(11:12:26) thl: tibbs, seems that's at least considered by the authors 
(11:12:31) thl: okay, accepted
(11:12:32) warren: Haven't been following kmod lately, will follow
whatever thl and scop decides.
(11:12:49) thl: warren, then you'll have a problem now
(11:12:57) thl has changed the topic to: FESCo Meeting in progress --
review kmod's -- zaptel-kmod
(11:13:01) thl: I stick to my -1
(11:13:07) thl: warren, scop voted +1
(11:13:20) thl: warren,
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=177583#c41 for
(11:13:38) thl: jwb, jeremy, I'd be very interested in your opinions on
(11:14:18) tibbs: I'm +1; frankly I wish someone would just take it over
and submit it upstream.
(11:14:24) thl: the whole thing in short: upstream doesn't want to get
the module included in the upstream kernel from kernel.org
(11:14:42) thl: and we only agreed to have kmod in extras
(11:14:52) jeremy: has anyone asked if they, similarly, don't want the
module included in Fedora?
(11:14:52) c4chris: Yea, the zero chance of it going upstream bugs me
(11:14:54) ***nirik had hopes that openpbx would do somethng like that,
but they seem stalled out these days. ;( 
(11:14:55) thl: if we make sure that the stuff we include get's upstream
(11:15:22) nirik: jeremy: that statement in the bug is from a digium
(11:15:22) tibbs: sysprof will never get upstream either.
(11:15:23) mjk_ is now known as mjk-
(11:15:43) thl: tibbs, but the author at least works close with upstream
(11:15:49) c4chris: tibbs, but sysprof is a temporary thing untill
oprofile is used
(11:16:01) thl: c4chris, agreed, too
(11:16:02) tibbs: Who's to say the Digium folks don't work closely with
(11:16:05) dgilmore: i use the module regullary 
(11:16:11) ixs: jeremy: the digium guy said something along: "we'd be
glad to help in a reasonable way to get it into fedora"
(11:16:20) nirik: "we'd like to see Asterisk and Zaptel in Fedora Extras
as well, so we'll do anything that's within reason to help achieve that
goal :-)"
(11:16:22) dgilmore: but im not really sure if we want it in extras
(11:16:38) tibbs: c4chris: My reading of the discussion is that the
sysprof author has no interesting in making his code work with oprofile
because it's too much work.
(11:17:07) dgilmore: nirik: where ia am at  cluecon  has some openpbx
guys  here  they nearly  have zaptel s need removed
(11:17:10) tibbs: He certainly hasn't made any progress towards that end
in over a year.
(11:17:19) ***jeremy really thinks that the answer for the zaptel stuff
is some cluebat'ing about usptream
(11:17:23) jeremy: maybe send dwmw2 to them ;-)
(11:17:26) nirik: dgilmore: cool. ;) 
(11:17:42) c4chris: tibbs, hmmm
(11:17:43) thl: jeremy, I had hoped dwmw2 would join the discussion on
the bug ;-)
(11:17:46) jwb: doesn't dwmw2 actually use this?
(11:17:55) dgilmore: jwb: yeah he does
(11:17:59) jwb: thought so
(11:18:03) jwb: he's in .tw now though
(11:18:08) nirik: dgilmore: then perhaps we should reject
asterisk/zaptel and wait for openpbx. ;) 
(11:18:18) thl: jwb, dwmw2 opinions is also here:
(11:18:34) dgilmore: i think im  going to go +1  just  beacuse i think
it will be really useful.  but i think longterm  we need to presure  for
it to go upstream
(11:18:39) jcollie[work]: don't hold your breath waiting for openpbx
(11:18:44) jwb: thl, ok.  i'll review those soon
(11:18:59) dgilmore: nirik: they say an openpbx  tarball will be a month
or more away from a 0.2 release
(11:19:10) thl: we probably really should shift this discussion to the
(11:19:43) thl: seems I'm the idiot that has to do this :-/
(11:19:46) c4chris: thl, yea I'm pretty undecided ATM
(11:19:46) tibbs: dwmw2's opinion is interesting but unfortunately isn't
selectively applicable to this module.
(11:19:51) thl: I'll try to find time for it
(11:19:53) abadger1999: This discussion could be generalized as "What
are the criteria FESCo should be applying?"
(11:20:10) nirik: thl: if you want I can post about it, since I was
going to review the kmod... ?
(11:20:16) c4chris: abadger1999, that would be useful too
(11:20:18) tibbs: abadger1999: Exactly my opinion as voiced on the list.
(11:20:21) thl: nirik, k, that would be helpfull
(11:20:47) thl: abadger1999, c4chris, tibbs, we work on that after that
(11:20:51) nirik: sure, can do no problem... 
(11:20:55) thl: nirik, could you use fedora-devel please
(11:21:05) thl: that seems the proper place for this discussion IMHO
(11:21:05) nirik: not extras? ok...
(11:21:09) c4chris: gotta run RSN...
(11:21:36) thl has changed the topic to: FESCo Meeting in progress --
warren: CVS Branch Request Process
(11:21:38) thl: warren ?
(11:21:42) warren: A simple improvement to the CVS branching request
process that requires no infrastructure changes is to append  the bug
number after package name.  So it would like * FC-4 FC-5 cluestick
284327   This is only an improvement to this interim process until we
can better automate the workflow in the package database.
(11:21:57) warren: that is all
(11:22:02) tibbs: +many
(11:22:07) mmcgrath: +1
(11:22:11) c4chris: +1
(11:22:15) bpepple: +1
(11:22:23) abadger1999: +1
(11:22:24) thl: +1
(11:22:34) dgilmore: +1
(11:22:45) thl: k, that was easy ;-)
(11:22:50) tibbs: Frankly I don't even think you needed to ask.
(11:23:03) thl: tibbs, agreed
(11:23:10) tibbs: warren: What about blocker bugs for branck requests?
(11:23:14) thl: we don#t need to discuss each and every details
(11:23:26) tibbs: NEEDBRANCH-FE4?
(11:23:28) thl: just doing it until somebody yells often is okay for me
(11:23:29) warren: tibbs, doesn't handle re-requesting of new branches
for things already previously branched
(11:23:31) warren: oh
(11:23:32) warren: hmm
(11:23:50) warren: tibbs, I'd prefer this system, it is supposed to be
temporary anyway
(11:24:10) tibbs: It's your process to make as far as I'm concerned.
(11:24:14) thl has changed the topic to: FESCo Meeting in progress --
abadger1999 ,  Future FESCo elections
(11:24:17) thl: any news?
(11:24:23) thl: or shall we just skip it?
(11:24:27) abadger1999: Not enough time this week.
(11:24:29) abadger1999: skip it.
(11:24:31) maners [n=maners] entered the room.
(11:24:35) warren: wah!/
(11:24:43) warren: spot made the CVS examples into an image
(11:24:43) thl has changed the topic to: FESCo Meeting in progress --
c4chris ,  Package Database
(11:24:45) warren: heh
(11:24:57) thl: ohh, he left
(11:25:04) thl: warren, seems c4chris is interesed in this
(11:25:05) c4chris: not quite
(11:25:11) thl: ohh :)
(11:25:16) c4chris: but RSN
(11:25:28) thl: c4chris, so how to get this thing rolling?
(11:25:29) c4chris: can we discuss this next week ?
(11:25:34) thl: c4chris, sure
(11:25:38) tibbs: Lots of folks are interested in this, I think.
(11:25:54) thl: yeah
(11:25:55) c4chris: thl, thanks
(11:26:07) thl has changed the topic to: FESCo Meeting in progress --
tibbs , Encourage Extras reviews
(11:26:15) thl: tibbs, that's still on the schedule
(11:26:19) ***c4chris goes afk now.  Cheers and TTYL
(11:26:20) thl: what do we do with it?
(11:26:26) tibbs: Yes, low priority.
(11:26:27) thl: c4chris, bye
(11:26:46) tibbs: At this point not much is going to happen unless
someone has new ideas.
(11:26:58) tibbs: I think we're actually doing pretty good lately.
(11:27:05) thl: tibbs, then I'll knock it of the schedule for now
(11:27:06) ***bpepple agrees.
(11:27:06) tibbs: Several new active reviewers and such.
(11:27:14) thl: we can re-add it later at any time
(11:27:22) tibbs: If we could get devimgrunduz sponsored I think we'd
have another.
(11:27:47) thl: also co-maintainership might help a bit in the longer
term (at least if I get it workin as I'd like it)
(11:28:10) abadger1999: tibbs: He's more interested in maintaining than
(11:28:10) tibbs: If co-maintainership opens up a second enterance to
the process then I agree it would help.
(11:28:14) bpepple: thl: Yeah, that would solve the problem with people
only submitting one package.
(11:28:25) thl has changed the topic to: FESCo Meeting in progress --
Free discussion around extras
(11:28:26) abadger1999: But I'm close to sponsoring him. 
(11:28:34) tibbs: abadger1999: Lots of people are, but they still do the
odd review.  Every one helps.
(11:28:43) thl: anything else that needs to be discussed?
(11:29:22) abadger1999: Something of a side note -- mono may move to
%{_libdir} for FC6.
(11:29:34) tibbs: Wow.
(11:29:46) tibbs: Who pushed that through?
(11:30:03) abadger1999: We're trying not to do that (big change) but
there may not be any other choice.
(11:30:26) abadger1999: alexl is on vacation this week so no progress
until next week.
(11:30:38) tibbs: There shouldn't be any other choice.  It's terrible
the way it is.
(11:30:49) ***thl still hopes we can get gutenprint into FC6
(11:30:55) abadger1999:
(11:31:10) tibbs: thl: I think that's pushing it.
(11:31:37) thl: tibbs, ?
(11:31:54) tibbs: abadger1999: Heh. Yes, that bug kind of sums up the
(11:31:58) warren: thl, FC6 feature freeze was weeks ago
(11:32:02) tibbs: thl: it's probably too late for gutenprint.
(11:32:15) abadger1999: tibbs: Right-o *grin*
(11:32:16) thl: warren, I know, I know
(11:32:40) thl: we'll seems we can close the meeting
(11:32:42) tibbs: But if it's packaged well enough couldn't it just slot
right in?  Or does it conflict with core in some way?
(11:32:52) ***thl will close the meeting in 30 seconds
(11:33:09) thl: warren, really, gutenprint will imnprove hardware
support a lot
(11:33:23) thl: the new version supports a lot of printers gimp-print
(11:33:24) warren: thl, was this discussed on f-d-l?
(11:33:28) ***thl will close the meeting in 15 seconds
(11:33:32) thl: warren, nope
(11:33:35) thl: once on testing
(11:33:54) thl: warren, seems the biggest problems is: twaught lacks
hardware to test
(11:34:01) thl: (and of course time to work on it)
(11:34:06) ***thl will close the meeting in 5 seconds
(11:34:12) thl: -- MARK -- Meeting End

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]