FESCo-2007-09-20

bpepple has changed the topic to: FESCo meeting -- Meeting rules at http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Init process
f13here
* abadger1999 waves from the bleachers
* warren here
notting is here
nirik is here.
c4chris is here
bpeppleFESCo meeting ping -- bpepple, caillon, c4chris, dgilmore, dwmw2, f13, jeremy, jwb, notting, spot, nirik, tibbs, warren
* jeremy is here
tibbs here, barely.
bpeppleHi, everyone.
* spot is here
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- revisiting status of features which we gave a pass to go in post-feature freeze - jeremy
f13I'm here.
warrenDid everyone vote on the proposal?
oh, wrong topic
bpepplewarren: we can switch if you'd like.
--- bpepple has changed the topic to: -- MISC -- release team proposal
f13this was the proposal sent to the list.
bpeppledid every one get a chance to read the proposal?  And if so,  is anyone against it?
* c4chris wonder about point 3 ...
warrenIt is very short, read it if you haven't yet.
tibbsSorry, I'm very behind on mail at this point.  Anyone have a link?
f13you can also see it in wiki form at http://fedoraproject.org/wiki/ReleaseEngineering/Meetings/2007-sep-17#head-7280e101faed94c9aa87f590f920a38c2a9a66a8
nirikas long as dist tags will do the right thing I think it's fine.
warrentibbs, it is on fesco list, no links are possible
tibbsOh, that one.  I think I did see that.
bpepple+1, I'm fine with the proposal.
f13ugh, I can't get into the archives becuase I didn't actaully sign up for the list myself so I don't know what the password is.
spoti voted for it at the rel-eng meeting, so i'm an obvious +1
notting+1
f13bp, jkatz, kevin all voted +1
warrenI'm changing the part about the Orbital Laser to something more formal.
f13k
warrenThen posting it to fedora-devel-announce
c4chris+1 (but I think point 3 is more trouble than it is worth...)
bpepplehttps://www.redhat.com/mailman/private/fedora-extras-steering/2007-September/msg00021.html
tibbsIt seems fine to me, but I think we can expect folks to get confused about process.
warrentibbs, which part of the process?
jeremytibbs: I don't know how we avoid people getting confused about process unless we just don't do freezes
which is a little impractical :)
tibbsI certainly needed something like this during the F7 freeze.
f13tibbs: as confused as the 'mail releng to get your build in' process, or the huge question of 'is my build in or not' or 'i built for rawhide, why isn't it in rawhide' confusion?
bpeppleIs anyone against the proposal?
if not, I think we can consider the proposal approved.
f13hurray.
warrenOK, I'll send it off.
bpepplewarren: great, thanks.
--- bpepple has changed the topic to: FESCO-Meeting -- MISC -- revisiting status of features which we gave a pass to go in post-feature freeze - jeremy
warren"We will loudly advertise expectations of acceptable changes during the freeze period."
jeremyso, we gave some features a bit of a pass on the feature freeze and I think it's worth going through the list again (... as painful as it is) to get a current idea of where they are
bpepplejeremy: I assume xulrunner is one of the features you want us to revisit?
jeremyxulrunner is one.  NetworkManager was built at some point in the past 24 hours, so it might be in shape
jeremyalthough there's some large gaps right now from some quick testing before lunch
nottingNM at the moment is having issues with dbus system activation and selinux
warrennotting, still?
nottingstill? you mean it's changed in the past two hours?
jeremynotting: + no vpn, not sure what's up with kde applet (haven't even looked to see yet),
let's just do a quick touch on each of the major outstanding ones?
c4chrissure
jeremyfirst up -- generic logos.  notting has made a lot of progress on this one
nirikso, we are just seeing how bad it is? or are we talking about pulling them as features?
jeremynirik: revisiting if they need to be pulled mostly
nottinggeneric-logos is waiting on review, and some cleanup of gdm theme packaging
jeremyso I think generic logos is in reasonable shape unless notting says otherwise
niriknotting: I can review it in the next few days if no one beats me to it.
bpeppleI might have some time tomorrow to look at it, also.
* dwmw2 appears. sorry, on phone
jeremytexlive -- it looks like the review is still somewhat stuck from quickly looking through the bug (242416)
nottingis it still stuck on legal?
jeremynot blocking FE-LEGAL
tibbsI thought the legal bits had mostly been cleaned up.
jeremyand comments imply some legal fixes ~ 2 weeks ago
nirikyeah, it looks to just be lacking a reviewer.
jeremyI don't really see how this still makes F8, but that's just me
f13seems rather late to me.
also, do we want to somewhat forbid this landing as a update to F8 like a week after we release?
bpepplef13: probably.
c4chrisI'd feel more comfortable making this an F9 feature
bpepplewhat do other FESCo members think.  Should we pull this for F8?
abadger1999Does the current tex package have the legal issues that texlive has cleaned up?
f13-1 for F8
abadger1999: that's a really good question :/
tibbsAlmost certainly.
abadger1999If so I'm for finding some way to get texlive in, painful as it is.
tibbsI think texlive is very important for F8, but I just don't know if it's practical.
jeremyabadger1999: at the risk of slipping the release?
warrenWhat exactly is the risk of letting it in now?
jeremywe have another release in six months.  if a feature doesn't make it for one release, there's always next time
this is the entire reason for doing time-based releases
rdieterstandard problem: texlive has *lots* of commenters, 0 reviewers. ):
abadger1999Hmm... review implies that the legally questionable stuff was just removed; no substitutions were necessary.
Maybe we could do the same to the current tex package...
* nirik sees that the "Contingency Plan" on the texlive feature is ""
spotto be fair, i haven't seen the texlive-package get respun off new upstream sources yet
rdieterif I get done with kde4-related gunk today, I'll try step in and help out review-wise.
abadger1999s/says/implies/
jeremyabadger1999: could be investigated.  but I think that trying to hold up F8 to get texlive shoved in is the wrong answer
niriktest3 freeze is tuesday?
jeremynirik: yep
abadger1999jeremy: Yeah.  I'm just thinking, if current tex is also "non-free", should we really not be shipping it?
In which case we're lacking in a tex package at all.
bpepple-1, I think it's pretty late in the game to try to get this in for F8.
nirikwell, I suppose it could get reviewed this weekend... but man, sometime things must be cut off.
notting-1 from me
dwmw2-1. It just isn't that urgent that we need to change it for F8.
jeremynirik: and what about making sure that it doesn't build anything else in the 4000 source packages we have that end up using tex in their build?
dwmw2where's the massive user-driven imperative?
abadger1999But if we can see that stripping the problematic stuff is doable, it seems to be safer.
nirikugh, yeah... ok, pull it.
abadger1999: what is problematic again about the current package?
* jeremy is +1 to pushing to an F9 feature, if it's not clear :)
f13(:
dwmw2jeremy: +1 to that.
c4chrisF8: -1; F9: +1
rdieternirik: upstream gone. no maintainance.
bpeppleok, doing a quick count I see seven votes to pushing this back to F9.
nirikyeah, F8: -1, F9: +1 here
dwmw2I've been poking at the Bluetooth bits. We have an selinux policy fix and I'm working on one or two fixed on the bluez-* side, and it looks good.
jeremypoelstra is out (new baby), so I'll do the feature page twiddling
bpepplejeremy: cool.
dwmw2so I'm happy with the status of that feature.
f13oh that's where he is (:
I probably saw that I just forgot.
abadger1999nirik: Looking for the post about it.
nirikhttps://www.redhat.com/archives/fedora-devel-list/2006-November/msg00195.html
jeremynext up -- new NetworkManager
dwmw2the one issue with claiming improved Bluetooth™ support is the people might ask where our certification is :)
f13dwmw2: so just call it "bluetooth" and not captalize it (:
or "support for bluetooth devices"
dwmw2:)
jeremyso new NetworkManager was built early this morning and will be in rawhide when it meanders its way out
nottingi got your certification right here. from arkham asylum.
f13it has some... issues
jeremythere's definitely some work to be done there, but the issues are supposed to be resolvable in relatively short order
f13but a good chunk of progress has been made
If this were the state of it at test2 time, this would be a slam dunk in.
jeremyif we didn't have the live images (where it's on by default), I'd be substantially less worried.  but since it is, it could be... tricky
I think dcbw was going to sleep, though, so I haven't had a chance to talk with him about what I saw before lunch
f13as it is, I'm still pretty OK, but reserve the right to disable it by default if things don't look good later on.
nirikso the feature was for NM on by default and handling all network connections?
bpepplef13: +1
jeremynirik: I don't think we're doing on by default for everything
f13no?
probably way too late for that I suppose
jeremy(that was the original idea, but I don't see things being there.  for one thing, doing it sanely requires installer changes and there's no way in hell we should do those at this point)
nirikgood. I'm fine with it improving, but I am very very concerend with the lack of testing for all the non wireless/laptop cases.
f13wish we could make it more obvious to enable after install.
jeremy: gotcha.
maybe just maybe we can make it more discoverable post-install?
nirikI think pushing on by default for f9 might be ok... but not f8.
jeremyf13: would be string changes to do so :)
* jeremy is still worried about it given the high profile it has with the live images, but, feels like we're in between a bit of a rock and a hard place
nirikwell, the only alternative would be to backout to and older version?
bpepplejeremy: yeah, but I don't think it's something we can pull.
f13jeremy: I'm a bit happier knowing that we can somewhat have dcbw to fix bugs between now and release, which is better than before where we might maybe get a partial slice of his time.
warrenAre we talking about NM by default?
jeremynirik: yes, with other bugs which have been "will be fixed in 0.7.0"
nirikyeah, I think we run with it and hope it can get stablized.
jeremybpepple: I'm not disagreeing.  I just hope we're not shooting ourselves in the foot :-/
warrenIs NM by default usable AT ALL for servers?  (Can you specify static IP's or configure multiple NIC's?)
jeremybut, not like I haven't had to fix NM for things in the past ;)
bpepplejeremy: agreed.
jeremywarren: please scroll back and actually read the discussion.
nirik-lots for "on by default" at this point from me.
f13warren: we've been using NM by default on Live images since pretty much day one
warren: but we're not going to do NM by default for F8.
warrenOK, just wanted that to be clear.
c4chrisI'm ok with letting this in
f13I would feel kind of bad after dcbw pulled an all nighter to land this to say no now.
nirikok, so any other features? or votes on this one?
jeremyf13: need to sit down and really catalog the problems, though.  and make sure that we're tracking them
other outstanding feature is xulrunner
dwmw2I'm kind of lost on what the result of a 'no' vote would be. We're already said it won't be _default_, right? If I say -1 does that mean we don't even pull in the new build of the package?
f13jeremy: yes, after meeting I'm doing to do a fresh reboot with updates snaked out of koji-static and see where we're at with NM
dwmw2: correct, a -1 would be rollback to old version
as lame as that is, but it would protect against the REGRESSION monsters.
dwmw2the new NM isn't likely to be _that_ bad, is it?
c4chrisxulrunner looks pretty red in the feature page as it is...
f13dwmw2: it's a rewrite
dwmw2+1 for it from me, I think.
jeremydwmw2: it's basically a rewrite
dwmw2NM needed a rewrite :)
f13dwmw2: until last night wireless didn't work for encrypted networks
dwmw2you just make me happier about it
f13dwmw2: it's still missing important UI stuff, like being able to launch a vpnc connection
dwmw2hm.
jeremyso xulrunner was not building on ppc, dwmw2 fixed that and now there's a new build problem with it
f13also it dropped the need for dhcdbd which is reason enough for me to let it in.
jeremy: that's fixed by a glibc build.
dwmw2jeremy: you mean the C++ FORTI... that one?
f13jeremy: it was a glibc bug being tickled by xul.
jeremydwmw2: yeah
dwmw2gcc bug, surely?
f13dwmw2: glibc
dwmw2that one, anyway
f13dwmw2: jakub tracked it down yesterday
jeremyso now that that's fixed, it should be buildable I think.  and desktop folks were going to fix things not to clobber firefox and then start rebuilding things
dwmw2not  http://gcc.gnu.org/PR33506 ?
jeremyat least, that was the plan during the desktop sig meeting yesterday
nirikjeremy: can those rebuilds get done before tuesday? I guess again we hope so.
f13dwmw2: hrm, odd that he claimed it was glibc then filed a gcc bug.
* bpepple hopes so also, since liferea is one of those effected.
f1317:54 <@jakub> roland: actually, it is a glibc bug
spotnew gcc hasn't gone into rawhide yet
f13guess there was more dialog on that later.
jeremynirik: once xulrunner is built, it shouldn't be a big problem to do the rebuilds
dwmw2f13: that postdates the mail I'm looking at. Anyway, it's Jakub's Fault™
* f13 thinks that the re-opened rawhide after test3 would be helpful for all the xul rebuilds (if we do them)
jeremynirik: as most upstreams have switched to xulrunner a while ago
f13dwmw2: hah, yeah, I'm seeing more dialog in the backlog where ti was clarified.
nirikok, cool.
nottingjeremy: give them a while, they can switch to webkit
f13I hate to think that we're going to have such a switch like that /after/ test3 though :/
jeremynotting: heh
f13wwoods will have to get his test on.
jeremyf13: if things aren't done by freeze time next week, then we have a harder conversation.  as I don't really want to see it post-test3 either
bpepplejeremy: +1
jeremy: any other features we need to discuss?
jeremythat's the list of ones that I think were outstanding as of test2
as far as being "testable". probably need to do a poke of feature owners just to make sure they update with current status for test3
f13jeremy: so you want a fixed xul to land, and all apps rebuilt, before Tuesday?
jeremyf13: yes.  and a pony.  :-)
* f13 thinks somebody is going to have to start yelling /loudly/ at the xul consumers.
dwmw2NO MORE BLOODY PONIES
wwoodsxulrunner testing? yeah, that one's not too bad
spotdwmw2: if your pony is bloody, you're riding it wrong.
c4chrisbut we love ponies
f13I'm a little disgruntled with CodecBuddy (aka codeina)
jeremyf13: it's not that large of a list, really.  and "we" can help  (collective "we" of rel-eng, cvsadmins, fesco, etc)
f13but...
* spot is rebuilding gcc now
wwoodsit's pretty obvious when apps are using xulrunner, and it's usually fairly obvious where the problem is when xulrunner is being a problem
dwmw2spot: ah, maybe so
spotgoing to work on firefox as soon as it is done
f13spot: beware the wrath of the tools folks.
wwoodserr.. yeah. you know what I mean. if xulrunner acts up it's easier to spot than, say, figuring out what's misbehaving if you can't play music
spotf13: locally. not through koji.
f13spot: I ssee.
dwmw2hm, whoever imported the ppc64 patch into xulrunner imported the wrong one. The one I posted to the mailing list was correct, not the one from mozilla bugzilla.
spoti think caillon did that before he went on vacation
so he probably didn't see the mailing list post
bpeppleok, it looks like we've got about 10 minutes left.  anything else about Features? Or should we move on?
dwmw2no, he merged the even older firefox-1.5 patch :)
jeremydwmw2: I'll fix it up so you can commit the right patch ;)
dwmw2jeremy: really? That would be spiffing :)
nirikif codecbuddy isn't matching the listed features, should we ping them to change it, or pull it as a feature?
f13honestly it looks like it's only missing oen big thing, that it be driven to a webpage instead of a local popup.
the local popup is right now all hardcoded for fluendo stuff, but since we don't have another vendor that's not too terrible.
dwmw2I disappeared during last week's meeting. Was there anything I was supposed to do following the kernel module thing?
f13I'dl ike to see it a bit more Fedora branded and made clear that Fluendo is just an option (the only option right now) but that it isn't just a big commercial for Fluendo
c4chrisf13: agreed
nirikok, so file bugs and see if they will just fix things? That sounds like the good next step to me.
f13but that can be fixed between now and release.
bpeppledwmw2: I think the only thing we talked about after you left was the wiki changes that needed to be made.
f13nod
* notting has some ideas, but no time
--- bpepple has changed the topic to: FESCo meeting -- Free discussion around Fedora
bpeppleok, we've got about 5 minutes left.  anything we need to discuss before wrapping up this week?
wwoodsThe fact that Fluendo is the *only* existing legal option needs to be kept in mind during any codeina discussion, I think
jeremywwoods: you forgot "in the US"
bpepplejeremy: ;)
f13wwoods: yes, but that /could/ change and we should be sure to have a flexible app to handle that.
jeremywwoods: other countries have different legal restrictions
wwoodswe're in the US, though. fluendo is the only thing we can legally point to right now, isn't it?
* rdieter thinks so...
c4chrisI don't think you'd be forbidden to mention that things can be different in other countries...
jeremywwoods: some readings of the magic 8 ball say "ask again later"
wwoodsso the fact that the app only points to fluendo right now is troubling
but there is *no alternative at this time*
jeremybut yeah, I'm not that bent out of shape about it at the moment
nottingnote that this is also being brought up by the board
wwoodsI'm just trying to make sure that further (unavoidable) discussion of codeina reflects that fact
f13I've got nothing further for FESCo at this time.
wwoodsI'll probably be tracking down feature owners about testing info
bpepplewwoods: cool.
f13goot.
wwoodsesp. pulseaudio. that one worries me mightily
abadger1999So tetex does have pieceswith no license inside the tarball.
wwoodsit's so cool when it works but so hard to debug when it doesn't.
f13speaking of, we have compiz-fusion now, but still no tool to fiddle with it.
dwmw2wwoods: feel free to hassle me about bluetooth testing
abadger1999there's also at least one thing that has non-commercial use in the file
* dwmw2 tries to find the firefox window in which he was half-way through buying a set of bluetooth headphones
nirikf13: are there more buildid rebuilds needed before test3?
wwoodsdwmw2: yeah, do you have some recommended tasks? I've got a headset and a phone but that's it
nirikalso, what about packages without the new license tags?
wwoods(I really want to do official support for the wiimote but that's outside of this discussion)
dwmw2pretty GUI stuff should let you browse the phone's files.
we don't let have PPP support -- that's waiting on NM development. Although it's easy enough to set up
wwoodsdwmw2: right, but what's the procedure for setting up the machine for testing? Should bluetooth just magically be enabled on a machine with proper hardware?
dwmw2I use modemlights_applet for GPRS "dialup"
wwoodswe can discuss outside this meeting
f13nirik: yes
wwoodsbut yeah
wwoodsthose are the sorts of questions
bpeppleok, unless there is anything else, we can probably wrap up the meeting.
dwmw2yes, I think it should. If not, hassle me and/or hadess
f13nirik: sending up a new list today
dwmw2yeah, let's do it later. SIGWENCH imminent
* bpepple will end the meeting in 60
bpepple will end the meeting in 30
wwoods(SIGLUNCH?)
* bpepple will end the meeting in 15
dwmw2that too. Well, dinner.
bpepple-- MARK -- Meeting End

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