bpeppletopic FESCo meeting -- Meeting rules at http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Init process
--- bpepple has changed the topic to: FESCo meeting -- Meeting rules at http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Init process
* warren here
* poelcat here
tibbs|h here
* nirik is around.
rdieter sits in the rabble-seats, shares popcorn with knurd
c4chris is here
* jima here (rabble)
bpeppleFESCo meeting ping -- bpepple, caillon, c4chris, dgilmore, dwmw2, f13, jeremy, jwb, notting, spot, nirik, tibbs, warren
* dgilmore is here
jwbbpepple, will be a few minutes late
* knurd hands the coke to rdieter
bpeppleHi, everyone; who's around?
jwb: np.
* kasal here, willing to join the discussion on "packages in buildroot'
jeremybbSpot and f13 will be late too
bpeppleThat's fine, we can start off slow.
* dwmw2 is good at slow
--- bpepple has changed the topic to: FESCo meeting -- Any objection to this week's report from FPC at https://www.redhat.com/archives/fedora-maintainers/2007-August/msg00702.html
bpeppleEveryone get a chance to read Tibb's summary?
tibbs|hIt's just a few minor changes, really.
* bpepple read it, and didn't have any objections to the FPC's proposals.
tibbs|hTweaks to the PHP guidelines to come up to date with what new PHP stuff is doing.
jeremybbI don't remember seeing anything controversial
tibbs|hAnd a clarification which probably doesn't need to come before FESCo anyway.
* nirik has no problem with it.
warrensorry, my IRC proxy died
* warren here
tibbs|hNext week's python eggs proposal is more interesting, I think, so folks should weigh in on that.
If they care, of course.
bpeppleok, I don't see any objections on the mailing list or here, so I think we can safely move on.
* jwb returns
dgilmorelooks fine to me
c4chrisfine for me
bpepplealright, movin' on........
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- extending mandate (permanent?) for EPEL - nirik
bpeppleThis seems like a no-brainer item to me.
jwbfine with me
dgilmorei think we need to decide  elect some people or apoint new people
nirikyeah, I don't recall why it was limited...
warrenthis really needs discussion?
jeremybbJust do it... Good work is being done, there seems to be a growing user base which will lead to more contributors too
dgilmorenirik: it was limited to see if it worked  and was really needed
tibbs|hIt would need discussion if EPEL hadn't actually done anything useful.
jwbdgilmore, i think we can leave that up to the EPEL committee...
knurddgilmore, does FESCO need to/want to decide that?
nirikdgilmore: yeah, sounds right I guess.
dgilmoreknurd: i kinda think fesco should say epel needs to have some elected people
caillonjust got in, the guys at thai jasmin were slow today
dgilmorebut we have a had a couple of elections recently
nirikI like the idea of EPEL being more SIG like personally... people just step up and do the work they want to/are able to... but I suppose we do need a way to decide difficult issues.
dgilmoreJust so everyone knows whatever happens i wont be involeved in EPEL to the level i have been anymore
jeremybbI care more about there's stuff getting done than elections or appintments for the sake of elections
bpepplejeremybb: +1
dgilmorenirik: i think SIG more than steering committie personally
nottingdgilmore: no time?
dgilmorenotting: yeah
i got all the piece in place which ate alot of time. i have other things i want to work on that need alot of my time
nirikWell, if it's SIG like it's more a matter for the SIG to decide how to manage itself I would think? unless FESCo wants to appoint someone to report to fesco from the sig or something.
dgilmore: thanks for all the hard work on EPEL stuff, BTW... it's really appreciated.
bpepplenirik: I'm all for EPEL deciding how to manage itself.  knurd meeting summaries are fine w/ me.
dgilmorenirik: SIG like the SIG decides who gives a report back to fesco
nirikok, sounds fine to me... so for now we have an OK from fesco to keep going... and manage as we see fit?
bpepplenirik: yup.
jwbas long as you don't go insane :)
nirikwell, we can try. ;)
bpepplenirik: anything else in regard to EPEL, or should we move on?
caillonplease to not be doing the insane thing, either
niriknothing in this topic I don't think... move on.
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- Review of new features status http://fedoraproject.org/wiki/Releases/8/FeatureList - poelcat, jwb
bpepplepoelcat: you wanted to talk about dropping features not finished.
jeremybbI think that needs to be on a feature by feature basis
bpepplejeremybb: correct.
dgilmorejeremybb: i.e.  is it nearly done  or no where near done
bpepplehmm, seems poelcat has stepped away.
poelcatbpepple: yes
poelcatbackground: https://www.redhat.com/archives/fedora-devel-list/2007-August/msg02087.html
jeremybbCould be even a little more involved
poelcatDoes FESCo have a preferred course of action?
jeremybb: it is
tibbsThings that aren't done still have some time to get done, but we need to triage those features which have no chance.
And status updates are mandatory for us to do that.
poelcattibbs: then what does "feature freeze" mean?
tibbsWe've chosen the features that will be in the release.
poelcatwe also said we would drop them if they weren't done on time :)
tibbsWhen has "feature freeze" ever meant anything else?
poelcat: if they weren't in a testable state.
jeremyfeature freeze means that "feature is largely to a reasonable place and is testable, bugs still to come".  _or_ that the feature has to get an okay to continue to go in
tibbsMy point is that if we can't even get status updates then the features should be dropped.
poelcathm... we didn't say that here: http://fedoraproject.org/wiki/Features/Policy#drop
f13they don't have to be 100% done, they just have to be in a testable state by test2, to be shored up throughout the rest.
dgilmorei think right now we need to cut what is not at least 75-80% feature complete
bpeppledgilmore: that sounds about right to me,
poelcati'm just trying to follow the policy we set in the beginning... naturally we can change it if we all agree :)
tibbsWell, "wakeups" probably gets a pass unless we can't honestly say that things are better than they used to be.
drago01dgilmore: what if they are >=75-80% but the page is not updated?
dgilmoredrago01: feature owners should be keeping the status updated.
nirikno updates should be == move to f9 target/dropped for f8
jwbdrago01, people got personal emails to update their page.  if they can't do that, it's their fault
drago01jwb: ok fair enough
bpepplepoelcat: I see your point, but using the 100% threshold, don't we have only 4 features done?
jeremythe fact that we drop a feature page from being tracked does absolutely nothing as far as "does code happen", though.  _that_ is why you can't just say "sorry, try again"
caillonjeremy: +1
poelcatbpepple: yes; so we need to tweak the policy or do what it says :)
poelcatwhich is how these things go
cailloni always thought the "Feature" list was mostly for marketing anyway
jwbcaillon, yes and no
cailloni didn't say all.  i said mostly
jwbcaillon, for example... presto could have been a feature.  but it wasn't in place, and now it won't be for f8 ever
jwbthe only controversial feature i see on this page that might get dropped is BigBoard
tibbsTake the "Dictionary" feature.  It's not done, because nobody wrote the aspell compatibility layer.  But firefox and openoffice can share dictionaries now.  Do we just not mention that as a feature?
jeremyjwb: except that the plugin is still available for people that want to install it.  and we're even still hoping to get to where the repos have deltarpms available.  so it's not black or white
poelcattibbs: or you change the description to reflect what is done
and propose the rest for F9
bpepplepoelcat: I think realistically we need to tweak our policy. I would lean towards modifying it to dgilomore's suggestion of 75-80%.
jwbjeremy, i meant in the fedora scope of things
jeremy, you aren't going to enable deltarpms in the official repos now, are you?
jeremyjwb: we talked about it last week.  and if we can get to where they're there, yes.  because without the plugin being installed, advertising it, etc then it provides a real value both for getting a leg up on testing and for users that are comfortable with it
dgilmorejwb: when they are available we could
jwbthat sort of pisses me off
but lets not focus on presto at the moment
nirikcan we agree to drop all features that are less than 50% and/or have no updates?
jwbnirik, +1
nirikor less than 75% / aren't testable
jwbyeah, something like that
poelcatnirik: "testable" is too ambiguous... how do we make that determination?
s/we/I :)
jeremynirik: what does "dropping" them mean?  I'm looking at the list of things that fall into that category and just not really seeing much for what it really means
f13poelcat: percentage done is also ambiguous and somewhat arbitrary isn't it?
tibbsxulrunner is one, I guess.
nirikpoelcat: yeah, thats true... approved to go on by fesco?
jeremytibbs: yeah, xulrunner is the only one that I can clearly understand what dropping means
poelcatf13: we don't have a field for "testable" :)
nirikjeremy: I would say that means they get moved to 'f9 feature' and wait for that cycle ?
jeremytibbs: and making a hunspell-based libaspell
tibbstexlive is another.
jwbdoes it occur to anyone else how ambiguous this whole thing is?
poelcatf13: and i can't tell just by looking at the feature page
jeremytibbs: texlive says it's at 95% :)
caillonxulrunner is on the feature list?
tibbsBut it still hasn't passed review.
jwbseems we were doing fine up until we had to really do anything about the features
f13"is it in a state that can be used by users and are problems they find useful to the development of this feature?"
f13if it's in a state that is know broken, and users using it will just cause noise, that's not "Testable"
poelcatcaillon: it got approved
jeremytibbs: yeah, I'm just saying that if we're following that criteria, it's not there
caillonoh i guess rahul submited it
drago01caillon: rahul added it .. can you remeber the flamewar on the mailing list?
caillondrago01: i'm in too many flamewars :)
notting_so, basically, we want to do a reality check and move things to 'not this release'?
drago01caillon: and you said "done in 2 weeks"
jeremybut changing the bookmarks is going to happen whether it's a "feature" or not, generic logos will continue to march a little bit forward probably, wakeup problems will get fixed as filed,  ...
bpepplenotting_: yeah.
caillonyeah, and then stuff happened.  it's sitting on my harddrive.  i have minor changes to the olpc one, and should just import that
c4chrisnotting_: I think yes
dgilmorespot: one feature thats not there but should show results.  probably soon after F-8 is done is secondary archs
poelcatf13: i agree with you... how do I collect this information in an efficient way for each feature?
dgilmoreopps sorry spot
jwbdgilmore, it's not a feature
dgilmorejwb: sure it is
caillonf13: what is the "test2" deadline?
dgilmorejwb: enabling many more archs is a feature
f13caillon: yesterday
jwbdgilmore, no, i meant it's not a Feature
f13or the day before really, but I did a somewhat silent pull of all yesterday's stuff.
jwbwe aren't tracking it as such
caillonany leeway there?
dgilmorejwb: thats what i said
one feature thats not there
* bpepple thinks we're getting off track.
jwbwe are
notting_it's an ugly process, but we could go through them one-by-one with the current state and do go/no-go
dgilmorenotting_: i think we should
f13caillon: some, I'm pretty sure we're going to slip anyway.
drago01there are many features that aren't feartures (untracked)
caillonbut what does no go mean?
notting_caillon: not-for-f8
f13caillon: no-go should mean don't advertise it for F8, and use the Contengency plan.
caillonnotting_: but does that mean we back out changes if they happen anyway?
dgilmorecaillon: that it wont be in the release initially and we dont push it as something new
bpepplecaillon: not for f8 as an advertised feature.
f13which is supposed to be required of each feature.
this is why I harped on having a Contengency plan.
notting_caillon: depends on the contingency plan, i suppose
f13said plan could simply be "the bits will still land, but we won't advertise nor make them the default"
caillonso let's do the 1by1 thing
f13this one doesn't have a clear indiciation of what the Feature would be, other than having the packages available.
jeremyand they're available
jwbthe packages are there and testab le
notting_so, update the % and move on?
poelcatnotting_: feature owner won't respond
tibbsErm, FeatureBigboard only talke about online desktop.
Oh, Bigboard is online desktop. Duh.
bpepplepoelcat: I can contact walters, and see if he will update the page.
poelcatbpepple: thanks
caillonsome people are just swamped with mail; so i don't think mail is the best way to contact everyone
notting_walters shows up on irc from time to time. undue harassment there may work
poelcatcaillon: I would do in person vists, but can't get approval to travel ;-)
bpepplecaillon: I was planning to contact him on irc first.
caillonyeah, that's what i'd do
ok, next
warrenPercentage of completion: ??%
that's helpful
caillon"Make things better" is hard to put a percentage on anyway
tibbsWell, what has been made better at this point?
nottinghowever there's a specific list of tasks in 'current status'. so we should be able to at least put a semi-arbitrary number on it
f13this one probably just needs a clarification at some point of "what things are better, and can be liste din release notes"
poelcatcaillon: then maybe it is not a clearly defined feature
drago01well some stuff is already in rawhide
jeremyI don't think that we're going to have an amazing bluetooth experience in f8 (if so, I haven't noticed it), so probably not a marketry type of feature
drago01like the "browse device" feature
bpepplejeremy: I agree.  this might be something to pull.
tibbsI mean, the kde bluetooth stuff went in at some point, but that isn't even mentioned in this feature.
bpepplehadess packages will still be available.
jeremythe packages are there, some things are better, but we can't say "Bluetooth Just Works"
dgilmoretibbs: indeed
c4chrisjeremy: agreed
bpeppleDoes anyone object to removing this as a feature?
f13not here.
tibbsno objections from me.
* nirik doesn't object.
caillondo it
tibbsIf the patches to firefox are in, I think enough has been done here to mention it.
nottingcaillon: approval of patch for branch yet?
bpeppletibbs: agreed.
niriktibbs: +1
c4christibbs: agreed
tibbsWould be nice to get the vim patch in as well, I guess.
f13probably need to list what all has been merged to share as the release notes item.
caillonnotting: no
nirikis that patch in rawhide? for firefox?
caillonbut i can push harder
nottingthe firefox patch is *not* merged. it's on ff trunk. not going into f8 unless it's approved for 2.0.x branch
(whoops, didn't mean to speak for caillon)
dgilmorei think we need to add a note  on weather or not this is available for gnome, xfce, kde apps also
poelcatcan the owners tune up the feature page? please? :) http://fedoraproject.org/wiki/Releases/FeatureDictionary#head-10016c3f31a18eb3cc0f93030682b129502f0aae
nirikdgilmore: yeah, that would be nice... not too many xfce apps need spelling tho.
nottingdgilmore: kde uses aspell :/
caillonnotting: and right now there's some legal question about whether mozilla can even ship it
nottingcaillon: ?
nirikso, looks like this feature might be cranked up to 75%ish?
nottingcaillon: license mishmash?
warrencaillon, why?
caillonnotting: the dictionaries are LGPL only.
dgilmorenotting: but just a note  saying it works on <blah> desktops. make it very clear for everyone
nottingcaillon: it's just data, but IANAL
caillonit doesn't really affect us, but it affects them, which affects u
I know
moving on
poelcatwhat is the decision?
tibbsSo if the only real change here is that the dictionaries have been split out of openoffice but only openoffice uses them, then I'm not sure we actually have something worth talking about.
bpeppletibbs: yeah, you're probably right.
tibbsWhich is sad, really.
nirikyeah. Agreed. even having vim might have added enough...
f13hey, it's only the what, third release we've tried to do this?
nottingso, i guess move to f9 as an advertised feature. we'd really need the aspell layer to get both kde and gnome proper
caillonafter other distros get it first
tibbsWell, we can't advertise it if we don't actually have it.
poelcatnext is http://fedoraproject.org/wiki/Releases/FeatureGenericLogos
tibbsAnd it sure looks like we have nothing to actually advertise.
nirikhumm.. vim does have it already doesn't it?
tibbsnirik: It's not really possible to tell from the feature page whether vim actually has it.
nirikoh, it's in the spec and a with_hunspell variable thats 0 by default...
jeremypoelcat: for this one, I think we conitnue to track it because we want to know what the holes are (and some of them will continue to get plugged as they'll be "fixes" which don't impact the normal case)
tibbsWe shouldn't need to look in the source to determine these things.
poelcattibbs: +1
niriktibbs: agreed.
nottingpoelcat: i'd like a minor extension if for no other reason than how we solve this is dependent on the branded art, which wasn't finalized until now
dgilmorenotting: i think thats fine
f13I'm with notting.
* nirik is ok with an extension on this... but it should be done asap...
f13We're doing other things to make derivitives easy with F8, this should get some extra time to get in.
f13(speaking of, I need to go find all the things that require fedora-release and fix them)
warrenf13, in what way?
tibbsI think there's enough done with laptop improvements to warrant mentioning.
f13warren: fedora-release now provides 'system-release'
f13warren: so that you can drop in replace it and not have something named "fedora" in your package set.
bpeppletibbs: +1
nirik+1 for laptop improvements. Hopefully it will keep getting better.
f13yeah, get a break down of what we improved, release note them, and move this feature over to F9 as we always want to make ongoing improvements.
tibbsThis really needs an update.
nirikI think work should keep going of course, but we should target f9 with the 'on by default'
dwmw2doesn't look like it's ready in time
dgilmoremy main question is can we make it work without a gui
drago01caillon: status?
nirikI'd like to see a lot of testing on something like this... there are a lot of cases here that would need to get handled right.
caillonso update is:
i fucked up
and thought that today was the freeze date
cailloni just now got off the phone with dcbw
he's about to build the bits we need
so it will be ready to go by tonight if we are slipping as jkeating alluded to
tibbsWhich bits are ready to go?
poelcatcaillon: can you update the wiki to reflect what will be going in?
warrenhow certain are you that the new bits work?
tibbsThe stuff under "Required"?
caillonthe required bits, the rest will be continued to be worked on
and i'm sure we'll find some bugs as this is new code
but we'll be looking to fix those up too
f13if anything for f8 we need t omake it easier to discover and turn on to enable.
caillonthe main points is that the public interfaces are done and not going to change
nirikso you think it's ready to enable by default? are their server/command line only tools ?
caillonwarren: we are going to import the version that does work
tibbsNobody's talking about removing the possibility of configuring things the "old way", right?
wwoodsI assume this means we'll be changing the startup number (rating? order?) for NetworkManager?
tibbsI need weird routing and stuff that I doubt networkmanager supports.
drago01tibbs: afaik no
nottingtibbs: probably not yet ;)
caillontibbs: like what?
tibbs: we have the interfaces to support it now
though not the UI yet
tibbsstuff like /etc/sysconfig/network-scripts/route-eth0
poelcatcan we vote +/- so we can move on?
caillontibbs: yeah we've got the general support for that
poelcatseems like a wiki page update would answer these questions
f13I give a +1 to a small extention, trying to get it in for Test2
caillon+1 for me obviously
tibbsIf that's possible, +1
notting+1. although we might need to do some dancing as to how/where it's used by default
nottingpoelcat: i updated this. we've done all the 'feature' work. now it's just bugfixing.
tibbsThe question here is whether we've done enough to mention.
nirik+1 for NM... hopefully it can be pounded on testing wise before release.
bpepplepoelcat: +1
nottingtibbs: that's probably a testing issue. if f7 was good enough by accident ... no
* c4chris has to leave early... later folks
bpepplec4chris: later.
caillonand i can get xulrunner in today and rebuild the stuff that depends on it
nirik+1 for fixwakeups... I think it's a good thing to tout, and changes will be ongoing.
poelcatcaillon: pls update the wiki page too
f13ok, so +1 for xul.
bpepple+1 xul here also.
tibbsAll we have to do is get the xulrunner package in, right?  We can rebuild the packages over time.
nirik+1... would be a very nice thing to have.
caillontibbs: i think so.  firefox will not build against it as-is really, though
but other things like epiphany and miro can
tibbsHrm.  Another patch we can't ship?  Or something else?
caillonmultilib fun
bpeppleliferea also can be built against it.
dwmw2caillon: I thought xulrunner made that easier?
bpepplepoelcat: any other items?
kasalpoelcat: what about TeXlive?
drago01caillon: does it still have versioned lib paths?
caillondwmw2: yes and no.
poelcatkasal: i was only covering features < 70% or questionable
caillondwmw2: do you (or others) have a solution for:  firefox requires the following directory structure to build:  sdk/include sdk/lib sdk/idl sdk/bin
poelcati will update http://fedoraproject.org/wiki/Releases/8/FeatureList
bpepplekasal: It's 95% done.
poelcat: thanks.
kasalpoelcat: ... but it was questioned earlier. thx
poelcatbased on today's meeting; for F9 we can tweak the policy to refelet what works best
dwmw2caillon: where is this 'sdk'? Not in %libdir?
tibbstexlive might as well be 0% done if it doesn't pass review.
* warren has to leave now
tibbsAnd if the licensing issues aren't cleared up, it will never pass review.
dwmw2sorry, not familar with the details of this at all. (So no, I don't have a handy solution :)
caillondwmw2: sdk is mostly headers and such so, it wants to live in %includedir
poelcattibbs: so what % should we put it at? the inention is to set reasonable expectations for what will be in the release
dwmw2caillon: but has a sdk/bin subdirectory... ?
caillonyes :)
and sdk/lib
which make things interesting
dwmw2that's that kind of thing we had to fix everywhere else for fhs complliance
tibbspoelcat: all I can say is to ping on the review ticket and try to get a status.
f13poelcat: so off my head of arbitrary, 75% == Testable.
dwmw2to put stuff in sharedir/libdir/bindir as appropriate
kasaltibbs: I was going to offer that I help w/ the review of TeXlive.  Is that the only problem w/ texlive you know of?
f13poelcat: higher than that is fixing buts that have been reported.  Lower is 'not ready for wide testing'.
bpepplef13: yeah, that sounds reasonable.
caillondwmw2: yeah, but that's the reason why i haven't gotten firefox working with it yet
dwmw2fair enough :)
caillonanyway, we can move on i guess
f13but we really should put up a grid somewhere of what we expect the percentages to mean.
bpepplepoelcat: anything else?
f13(and if we do that, do we really need percentages or can we just use human words?)
poelcatbpepple: thanks... i think that was only 16 minutes ;-)
tibbskasal: All sorts of things are swirling around that review.  Naming for tex packages, licenses, whether to split or not, the update package, etc.
bpepplepoelcat: yeah. ;)
tibbsThere's just a ton of stuff there.
dwmw2f13: don't say that. You scare me... and remind me of a conversation Karen had with a gym teacher about percentages... and how one particular test mark was percent out of 60...
--- bpepple has changed the topic to: FESCo meeting -- Free discussion around Fedora
dwmw2did we decide to skip the kmod thing again?
caillonsounds like a good candidate comic for xkcd
kasaltibbs: ok, so the pkg review bug is a good indicator what to do.  thx
jwbdgilmore, spot: how's the 2nd arch thing coming?
tibbskasal: BTW, https://bugzilla.redhat.com/show_bug.cgi?id=242416
nirikwell, we are 11min over already... ;(
dgilmorejwb: fixing some bugs in the tool chain for sparc right now
f13dwmw2: just haven't got there yet, but probably since I haven't seen any communication from |jef| regarding the dell method.
kasaltibbs: thcx
f13I had a couple subjects too :(
dgilmorejwb: the secondary arch daemon is pre-alpha right now
bpeppledwmw2: yeah, I'll try to schedule at the beginning of the meeting next week.
jwbdgilmore, did we ever figure out why spot though the needed rawhide to run the buildsys tools?
bpepplethere was quite a bit of talk about the buildroot, but I'm not sure if we have enough time to discuss it this week.
dwmw2I can post it back to the list as knurd requested... even though the proposal is just a cut and paste of what was already said on the list
dgilmorejwb: we need to have python-2.5 in the buildroot
jwbs/though the/thought he
dgilmorejwb: we dont need it to run
tibbsThis feature stuff always takes up all of our time.
bpeppledwmw2: that sounds good.
dgilmorejwb: it was probably bad communication on my part to spot
kasalbpepple: I wrote an article about the buildroot.  Can this be scheduled for next week?
dwmw2I'm really unconvinced that the feature stuff is worth the time we spend on it (says he fairly hypocritically, having been absent for three weeks)
jwbdgilmore, so we have a koji instance running on sparc then?
bpepplekasal: that sounds good to me.  I'll add it to the schedule for next week.
dgilmorejwb: sparc.koji.fedoraproject.org
nirikso will we need to deal with features next week? or since it's freeze time it's over for this cycle?
jwbdgilmore, is that box outside the colo?
f13FESCo can assign a sub-commitee to deal with features..
dgilmorejwb: once we get a few things built we will switch it on live
bpepplenirik: I don't think so. We really have a lot of other items we need to discuss.
kasalbpepple: ok, I'll remember to join this mtg next week.
dgilmorejwb: its at my house
f13The Feature Police
jwbdgilmore, awesome
poelcatf13: i thought that was my job :)
dgilmorejwb: secondary archs have to host themselves
jwbdgilmore, i know, i was just verifying.  this is very good news
nottingnirik: i don't think 'because we ran out of meeting time' is a proper reason to end discussion for the things at the end of the list. although not sure having another interim meeting is 'better'
kasalall: please look at http://fedoraproject.org/wiki/StepanKasal/Add_to_Exception_List
niriknotting: agreed, perhaps a special session? and/or continue on email?
nottingpoelcat: can you bring all the features we didn't go over today to e-mail?
or do we want a separate meeting?
bpepplenirik: yeah, we've done special sessions in the past.
poelcatnotting: we hit everything < 70%
nottingoh. ok. we're more done than i thought!
poelcatdoing this over email has not worked in the past
* kasal will be glad to discuss the Add_to_Exception_List with anyone interested during this week
nirik is fine with whatever, but right now needs to go work on real work(tm)
jeremy thinks we need to get the add to exception list thing out of the way
bpeppleyeah, we should wrap it up for this week.
jeremybecause it really is impacting people
f13I do too.
bpepplejeremy: agreed.
f13I'm willing to stick around if others are.
bpeppleI can stick around, also.
tibbsI'm off today, so not else to do.
cailloni can for a bit
kasalwell, I'm willing to discuss awk.  ;-)  If anyone is interested...
f13the proposed list doesn't look that bad.  I do want to do some comparisons between what is currently there, and what more will be brought in by this, and if it's at all possible to do a Fedora install (including rpm-build), log in, and /not/ have some of these things installed.
that's really the goals of the minimal build root.
A) prep the build root quickly, B) keep users from having to guess at what they're missing when they do a rpmbuild --rebuild foo.src.rpm
dgilmorei think that the assumption that deps would not change was a bad assumption made by many people.
f13and we did a poor job messaging that they could change.
(and an even poorer job of keeping the implicit list updated with changes)
f13especially since the changes aren't the same across the releases.
* jeremy says deputize f13 to look at the proposed list and take what makes sense within the constraints he said above
kasalf13: awk will pass the "minimal install test" by shere luck: initscripts require it.
dgilmoreso going forward i think awk is probably ok to add to the exception list
bpepplejeremy: +1
I +1 that.
dgilmorepossibly make sure sed is there  but thats it
f13If I make the change, it's going to be across the board.
nottingkasal: initscripts isn't in the build root atm in rawhide
dgilmoreim ok with f13 makeing a recomendation
wwoodssed's in coreutils, innit?
f13I do not want to have collection specific build groups.
we should talk about perl-devel though.
jeremywwoods: no, it has its own package
wwoodsoh ho. it isn't.
nirikif we re-add gawk, someone should go grep through all the packages that added it and ask them to take it out?
f13nirik: no.  Honestly extra BRs don't hurt
tibbsI personally don't think that gcc should be in the minimal buildroot, but what do I know.
dgilmoref13: i think that people should have to RB perl-devel
f13I do too, but I'm sure perl people will yell at me.
* notting thinks we should explicitly add sed, grep, awk, util-linux, remove perl
dgilmoretibbs: well we could use icc
f13perl-devel was added as a short term work around, and we never removed it.
tibbsI think that people should have to BR the actual perl modules they need.
nottingprobably findutils as well
tibbsWe agreed to remove it.
f13ok, I"m happy with that.
tibbsAnd is it util-linux or util-linux-ng these days?
jeremynotting: rpm-build requires findutils, but probably good to make it explicit
tibbsDoes -ng not provide the non-ng package?
kasaltibbs: util-linux-ng will provide util-linux, so if the list says util-linux, it's not that bad.
f13tibbs: it does.
as well as 'setarch' and the other htings that were their own packages prior to.
nottinginfo as suggested by kasal seems ok
f13now, if only I had a working rawhide to test installs with... (:
kasalActually, there was a suggestion to state that "if the package is in Exception List, you can omit the BuildRequires tag"
this formulation would prevent unwise depending on current state of deps.
f13yeah, you can, but not must.
tibbsYou can.  I used to ding people on this in reviews but experience has taught me not to.
kasalf13: sure you do not have to.
tibbs: so did I. But when I think about it, I see it was silly.
tibbsSome reviewers still want to see the BR list trimmed but as far as I'm concerned if the package builds in mock I don't really care what you have there as long as you're not pulling in hundreds of megs of packages you don't need.
f13anywho, I think that was enough general approval for the proposal, so long as it fits in teh criteria I've put forth, so I'll work on that hopefully today while waiting for things to compose.
bpepplef13: great.  thanks.
kasalf13: great!
bpeppleanything else, or should we call it quits?
f13yeah, please think on http://fedoraproject.org/wiki/LubomirKundrak/SecurityUpdateProcessDraft
kasalf13: when I mentioned the "new formulation" I meant something else:
jwbbpepple, could you mention the updated status of the secondary arch stuff in the meeting minutes?
bpepplejwb: yup.  I was planning on writing them later this afternoon at the coffee shop.
kasalf13: the old formulation pointed people to the full expansion of Exception List.  The new one should say: if a BuildRequires comes to your mind, and is not in the ExceptionList proper, do add it to the spec.
jwbbpepple, thanks.  i know several people are anxiously awaiting that news :)
bpepplef13: I'll make a mention of that in the meeting minutes also.
f13kasal: gotcha, yeah, that sounds right.
* bpepple will end the meeting in 60
bpepple will end the meeting in 30
kasalf13: that would mean we have to add binutils explicitly, too
* bpepple will end the meeting in 15
kasalf13: but not much else, I guess.  We do not want to mention fielsystem or some such.
bpepple-- MARK -- Meeting End
Thanks, everyone!

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