=== sladen [paul@starsky.19inch.net] has joined #ubuntu-meeting === Skif [~emschwar@216.17.176.4] has joined #ubuntu-meeting === Skif [~emschwar@216.17.176.4] has left #ubuntu-meeting ["Leaving"] === lamont [~lamont@mix.mmjgroup.com] has left #ubuntu-meeting ["Client] === pitti [~martin@195.227.105.180] has joined #ubuntu-meeting === Safari_Al [~tr@ppp139-87.lns1.adl2.internode.on.net] has joined #ubuntu-meeting === Safari_Al [~tr@ppp139-87.lns1.adl2.internode.on.net] has left #ubuntu-meeting ["Leaving"] === fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-meeting === pitti [~martin@195.227.105.180] has joined #ubuntu-meeting === pitti [~martin@195.227.105.180] has joined #ubuntu-meeting === tbw [asdfasdf@port14.cvx2-ejb.ppp.cybercity.dk] has joined #ubuntu-meeting === Keybuk [scott@descent.netsplit.com] has joined #ubuntu-meeting === ..[topic/#ubuntu-meeting:Keybuk] : Technical Board Meeting 1600UTC Today: http://wiki.ubuntu.com/TechnicalBoardAgenda === tbw [asdfasdf@port14.cvx2-ejb.ppp.cybercity.dk] has left #ubuntu-meeting [] === T-Bone [varenet@T-Bone.developer.debian] has joined #ubuntu-meeting === sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting [06:04] Hi sabdfl [06:04] (I'll be about 2 minutes) [06:04] hi everybody [06:05] hey everyone === silbs [~sbsm0084@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-meeting [06:05] mdz around? [06:06] sabdfl: yes, in #u-devel [06:06] just pinged him [06:06] who else do we need? kamion? [06:06] me too :-) [06:06] elmo perhaps? [06:07] Keybuk: ? === hhqw [asdfasdf@port9.cvx1-ejb.ppp.cybercity.dk] has joined #ubuntu-meeting === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-meeting === elmo [~james@83.216.141.215] has joined #ubuntu-meeting [06:08] Keybuk: ready? [06:08] yup === Keybuk is running massively behind today; haggai popped in for half an hour and stayed for four :) === sabdfl is running massively behind today. tla and postgres and make on macosx :-) [06:09] mdz: set? [06:09] yep [06:09] ok, warty status matt? [06:10] looks like no new critical bugs were opened while I slept, so currently we are still in release condition as far as bugs [06:10] Kamion: the archive-copier changes went in last night as well? [06:10] I haven't been able to check in with jdub since going to sleep last night, so I'm not sure about the status of his work [06:11] from my side, waiting only for new artowrk packages from jdub [06:11] lamont produced a working live CD at the data center [06:11] working? [06:11] I downloaded it and used it [06:11] network, apps? [06:11] mdz: I checked today's ppc image, archive-copier and stuff was okay [06:11] sabdfl: yeps.. tested here too [06:11] great [06:11] tested network connectivity, firefox, openoffice, etc. [06:11] for some reason my sound came up muted [06:11] wireless was detected and the driver loaded [06:12] if we have to tweak live cd, does it involve package tweaks or is it all behind the scenes? [06:12] fabbione: was your sound muted? [06:12] sabdfl: in this case, we had to remove something [06:12] mdz: yes, today's CD seems to work with it [06:12] mdz: i don't have any speakers yet sorry [06:12] mdz: they are still packed and they will be for a long while [06:12] sabdfl: the entire desktop set + the winfoss stuff was > 650M [06:12] just by a hair, but still over [06:12] reason for asking is if we can release tomorrow for install, and put out an rc livecd [06:12] lamont knows 3-4 issues (including sound) w/ live cd [06:12] sabdfl: if we could remove one of the winfoss apps, we could restore the font package we removed [06:13] lamont asked and i though we should drop a windows package rather than compromise warty [06:13] done [06:13] sabdfl: agreed, but I didn't want to do that without asking you [06:13] ok, let's drop that astronomical program [06:13] agreed [06:13] are the live cd issues fixable without package uploads? [06:13] in other words, can we consider the packages final for release? [06:14] are the fixes in the underlying layer? [06:14] I haven't caught up with lamont yet this morning, so I'm not sure what the issues are [06:14] I'll follow up after the meeting [06:14] 1) artwork will change [06:14] (16:20:56) lamont: 2) sound is busted some places (due to differences in hwdetection between morphix kernel and ubuntu...) [06:14] 3 was the space thing and 4 is 4) other hardware detection inconsistencies between live and install CD's exist. [06:14] 2) lamont said alex fixed === lamont [~lamont@mix.mmjgroup.com] has joined #ubuntu-meeting [06:15] don't think we can do anything further about hw detection inconsistencies [06:15] 4) is a known issue we must live with for warty [06:15] it's a hoary goal to make it consistent across the board === amu [~amu@195.71.9.198] has joined #ubuntu-meeting [06:15] we knew warty wouldn't have perfect consistency [06:15] lamont: you said alex fixed the order of sound device detection, right? [06:15] so i don't think should hold back the release [06:15] we can't call the packages final until we have final artwork, since that's in packaged form [06:15] mdz: he said he reveresed it, but didn't have a 2-sound system to check against. [06:16] lamont: give me a URL and I'll start downloading now [06:16] sound is still unhappy on my vaio [06:16] http://cdimage.ubuntu.com/sounder-test/live/warty-live-i386.iso.torrent [06:16] Or, if you can't use bit torrent, the actual iso is at: [06:16] http://cdimage.ubuntu.com/sounder-test/live/warty-live-i386.iso [06:16] ok, will test when I have it [06:16] building a new CD image +ttf-baekmuk -windoze.celestia [06:17] lamont: that still fits, right? [06:17] mdz: yes, but do the live cd changes also require package changes, or is it just the autobuilder that changes? [06:17] should. that's what I'm verifying [06:17] looks like the problem is desktop <-> laptop systems based, i tested it today on 4 differnt systems, on laptops it fails on desktops sound works [06:17] sabdfl: the remaining changes are artwork (which requires package changes AND build system changes, I believe) and the changes lamont is doing right now (which are purely build system changes) [06:17] sabdfl: sorry that was to mdz. liveCD changes require (maybe) syncing the local mirror, and then turning the crank again [06:18] mdz: artwork --> package change, sync new packages into the right place, turn crank [06:18] lamont: you don't need to drop the splash image into place? [06:18] or the grub image? [06:18] lamont: are you building locally now or in the LAN? [06:18] that's the package that has to get sync'ed/ [06:18] ok, but fixing sound, hw sync etc, that's an autobuilder change, not a package change? [06:18] Kamion: building home-edition right now, if it fits, I'll build the DC-edition [06:18] lamont: 'kay [06:19] "ubuntu, the colorado hills edition" [06:19] "live, from colorado" [06:19] fixing sound/hw detection is possibly a major overhaul of morphix [06:19] pass [06:19] worst case, sound is a user-obvious case of the hw detection not meshing. [06:19] unfortunately, users will judge Ubuntu hardware detection based on the live CD [06:20] we tried the trivial workaround for that case in the current CD. [06:20] even though they're completely different [06:20] mdz: very true [06:20] hoary will be better, right amu? [06:20] amu: in my case, it was simply that the sound devices were found in the opposite order [06:20] on my laptop [06:20] since it has a modem [06:20] ok, artwork [06:21] did anyone hear anything from jeff today? [06:21] for hoary, since we already have all the autodetection in our kernel/X/whatever, it probably makes sense to just roll the morphix patches over and include things that way, I would think... dunno. [06:21] i saw a note about him going to a meeting [06:21] sabdfl: i'm sure, with this manpower we reach asap world domination [06:21] 9am aussie time [06:21] but nothing fter that [06:21] btw, this liveCD has the latest grub screen from jdub [06:22] he was off to a presentation of some kind this morning (UK) [06:22] mdz: I didnt tested why, i tested more basic things, works/works not [06:22] amu: much of your list was "not for warty" stuff, sadly. [06:23] we either have to build a new set of packages based on what jdub last pushed out, or wait for him, or call him at home [06:23] lamont: ... i tested it from a view a user, which is imported for a user, in order to do his dayly work with it [06:23] i'd really like to have a final build tonight [06:23] we should call him [06:23] so we can get it mirrored ut [06:23] ok [06:23] it's only 0230 or so, no? :-) [06:23] descent scott% TZ=Australia/Sydney date [06:23] Wed Oct 20 02:23:51 EST 2004 [06:24] let's wake him up if he's asleep [06:24] amu: understood [06:24] he should be awake anyway :) [06:24] he did know that we wanted to do a build soon [06:24] Kamion: eheh [06:24] mdz: we can push stuff like jigit in post-CD-build, right? [06:24] sabdfl: should be possible [06:24] since it's Supported, not Base/Desktop/Ship [06:24] Kamion: into the archive? [06:24] no answer [06:24] yes [06:24] voicemail [06:24] mdz: yes [06:25] how will we deal with the flood of "please get xyz version x..z into universe, it fixes a ton of bugs" requests? [06:25] sabdfl: "use hoary"? [06:25] sabdfl: will sync for hoary, I imagine... [06:25] ok [06:25] are we intending to update universe? I assumed that would be frozen like supported [06:26] Keybuk: correct [06:26] freeze the whole thing [06:26] so yeah, mdz++ [06:26] we've been very flexible with updating universe stuff up until now [06:26] during hoary's freeze, can we setup a process for the community to nominate and approve syncs without affecting the core team? [06:26] but we do have to call an end to it [06:26] sabdfl: hmm, could work [06:26] for universe [06:26] the "universe team" [06:26] "masters of the universe" [06:26] as long as they are sane and sensible, they could, erm rule the universe [06:27] mdz++ [06:27] sabdfl: on the "should you care" front, much of multiverse is d-w j2sdk1.[34] === Keybuk mops the tea from his monitor [06:27] i am afraid that will give buildd (and lamont) a lot of headackes to unfuck * === sivang [~sg@80.179.93.130.forward.012.net.il] has joined #ubuntu-meeting [06:27] lamont: and in english? [06:27] howdy folks [06:27] lots of multiverse ain't built since we ain't got no java in multiverse [06:27] sabdfl: lacking java [06:27] oh wait. that was 'merican. [06:27] 'k [06:27] fabbione: build failures will be the responsibility of the universe guys [06:28] but warty remains warty. bug fixes for universe go into hoary. [06:28] packages: evolution, well it has some nice feature like spamchecks or groupware, outlookplugins, i think we should add those depends [06:28] mdz: i am more afraid of something that builds and kill the rest of universe [06:28] ok [06:28] just trying jdub's mobile [06:28] fabbione: also their problem [06:29] this is what it means for universe to be community supported [06:29] ok, let's move on [06:29] mdz: in theory it works perfectly.. let's see how it goes in reality :-) [06:29] Hoary starter plans [06:30] that was mine, I think [06:30] I estimate we have a solid few weeks of work for the review and merge process [06:30] *swaps hats* [06:30] Keybuk: hct status? [06:30] what I was going to offer, for the review/merge process was setting up a temporary arch repository [06:30] import debian from warty freeze to current day into that [06:30] that includes a lot of stuff that we can't really automate, even with hct [06:30] import warty into that [06:31] that'll give us an easy merge to create hoary [06:31] we need to check which bugs have been fixed, even if they were done in a different way [06:31] "easy merge"? [06:31] sabdfl: hct is good, but without any arch imports it's not useful [06:31] sabdfl: well, I can do an hct tablecloth-trick (jdub's name :p) on it [06:31] how many packages did we touch for warty? [06:31] fwiw I'm happy to merge d-i by hand; estimate about a week ... [06:31] sabdfl: many. many. many. [06:32] about 340 right? [06:32] sabdfl: at last count [06:32] that figure has only increased since [06:32] one option is to put the new things into hoary only [06:32] like x.org [06:32] gnome 2.9 [06:32] we can use hct to do the difficult bit, but as a general-use tool it's still waiting on those bloody imports :-( [06:32] 1018 source packages [06:32] stuff where we are really the head [06:32] 356 with ubuntu version numbers [06:33] so about 35% [06:33] and the number will only go up for hoary [06:33] what do we do with the rest of the packages? [06:33] (btw fantastic job warty team) [06:33] sabdfl: well, much stuff was sent upstream [06:33] pitti: yes, but we have no record of what was merged and what was not [06:33] I see more like 500 with ubuntu version numbers, including all components [06:33] pitti: yes, but a ton of stuff wasn't [06:33] and we have many changes which were not appropriate for upstream [06:33] sabdfl: for x.org i need xfree86 too. that's not really an option [06:33] no, unfortunately not in an automated way [06:33] elmo: I was only looking at main [06:34] fabbione: so manage the transition? [06:34] s/so/to/ [06:34] elmo has said it is no problem to set things up so that we automatically sync stuff that doesn't have an ubuntu version number [06:34] sabdfl: in terms of code we have 300K lines of patches from Xfree86 that needs to be reviewed to go into X.org [06:34] then we merge the remainder by hand [06:34] sabdfl: that's just "upstream" transition [06:34] is that likely to lead to real breakage in dependencies? [06:34] sabdfl: without taking into account all the packaging part [06:35] fabbione: daniel mentioned that 300k number as well; where does that come from? [06:35] since the packages we touched hardest are probably also most central? [06:35] there are less than 300k lines of patches in debian/patches [06:35] about half that [06:35] yeah, it's 492 including non-main components.. wonder why we ended up modifying so many non-main things [06:35] mdz: wc -l debian/patches/* [06:35] sabdfl: yes, there will be dependency breakage [06:35] elmo: libtiff transition in universe, etc. [06:35] ah, right [06:35] elmo: lamont had a build-dep mania one day [06:35] fabbione: diffstat debian/patches/* [06:35] ok... this isn't an area where i can lead guys, you know best [06:35] mdz: that will certainly result in much temporary breakage in d-i, but I think that's fine for hoary TBH [06:35] sabdfl: the issue is that we have a ton of work to do before hoary can really start to flow [06:35] sabdfl: which doesn't give us a break after the release at all [06:36] we have a semi-automatic solution that might work for everything we're not leading [06:36] it'll take a few days to make d-i work for hoary anyway, for one reason or another [06:36] team deserves a break [06:36] things like gnome we're leading Debian anyway === lamont must take a break post release, or find a new wife and a way to pay for the old one. [06:36] Keybuk: the semi-automation is not dependent on imports? [06:36] $ ls src/ubuntu | wc -l [06:36] 104 [06:36] Keybuk: what's the status on the imports? lifeless told me we were on our way [06:37] mdz: only package imports and I can do those and throw them away afterwards [06:37] sabdfl: well, arch.ubuntu.com hasn't changed in three weeks [06:37] guess about half of those belong to me to merge [06:37] lamont: lol [06:37] we should do a manual review up front, before trying anything fancy [06:38] ok, i'll chase lifeless on the arch merge front [06:38] we can probably eliminate 25-50% of those packages because everything has been merged upstream [06:38] have mostly been focused on malone [06:38] based on changelogs [06:38] mdz: hardest work will be stuff that's been branded, for sure [06:38] mdz: 1499 files changed, 161366 insertions(+), 67008 deletions(-) [06:38] Kamion: agreed [06:38] are our own patches clearly separated? [06:38] sabdfl: depends [06:38] sabdfl: only in a subset of cases [06:38] sabdfl: you can't clearly separate patches to .po files and stay sane at the same time [06:38] so are you guys READY for tla? [06:39] sabdfl: no [06:39] patches to .po files are evil :-( [06:39] Kamion: right [06:39] rather, tla is not ready for us [06:39] rosetta :-) [06:39] sabdfl: no for X.org [06:39] sabdfl: yes, but even so [06:39] tla isn't going to improve soon [06:39] we have an arch team drill for two weeks before es-conf [06:39] but that's jsut where we will set the goals for usability [06:39] it would be much less work to bring the patches forward manually than to get up to speed with tla [06:40] everyone would take 30 packages === Kamion is happy to use tla in a subset of cases, but wants to have a Plan B [06:40] the launchpad team has really battled with tla, and it's certianly not something i'm comfortable recommending for global FLOSS use [06:40] with any luck, 10 of those would be simple changelog review and revert to Debian [06:40] at the same time, it's proven fantastic when it works [06:40] sabdfl: did you read mako's blog entry? [06:40] nope, url? [06:40] that's not a huge amount of work for each of us [06:40] (er, when I say tla I mean raw/hct/whatever) [06:40] mdz: I think we could crunch through nearly everything in a week, personally [06:41] http://mako.yukidoke.org/copyrighteous/freesoftware/20041017-00.html [06:41] Kamion: we need to pace ourselves :-) [06:41] a week of fairly solid work, but I don't think it'll be terribly *hard* work [06:41] or planet.ubuntu.com [06:41] yes, very well put [06:41] Kamion: i agree..a lot of our local patches have been included in debian [06:41] the main issue is not that we can crunch through it, it's that sustaining the delta manually is just going to be "make work" without good tools [06:41] and the packages we've changed are all hopefully packages we've become familiar with [06:42] at least that's how it seems to me, but you guys would know better [06:42] mdz: pace? i wake up in the middle of the night yelling X.org :) [06:42] heh, I had a dream the other night in which everyone was a changeset and all trying to get through a single door [06:42] lamont: pay for the old one ? a easy solution, poison her, you go into prison and you have all day time for hacking [06:42] sabdfl: that is correct [06:42] but the tools simply aren't ready yet [06:42] sabdfl: that's true, but at the same time our feature goals for hoary are such that we need to have a place to put interim work on them fairly soon, without losing too much time with the tools [06:42] ok [06:42] seems to me bringing up the toolset would be better done in the middle of a release cycle, where it can be more easily parallelised? [06:42] we can create a simple source-package-merge tool which can automate the cases where there are no conflicts [06:43] that's not hard [06:43] we can set up a notification system whenever a new merge is needed [06:43] one option is just to ignore it till we freeze hoary [06:43] because at least then the targete isn't moving so fast [06:44] so, start work on all the new things for hoary [06:44] push upstream as much as possible [06:44] you mean, ignore merging new stuff from Debian? [06:44] no, other way around [06:44] mdz: Debian hasn't moved a *huge* amount yet [06:44] ignore bringing forward warty stuff [06:44] Keybuk: the RC bugs we've seen fly by say otherwise :-/ [06:44] sabdfl: I don't think we can do that [06:44] ok [06:45] that would revert all of our branding, all of our bug fixes, everything we've done to make things work right [06:45] the merge will only get more difficult as time goes on [06:45] i'm just wondering whether it's more work to port it forward and then keep it merged as debian updates before we freeze [06:45] the right time to do it is immediately when hoary opens [06:45] or bring it forward when we freeze [06:45] but then we have to merge every time the relevant debian package updates [06:46] in my experience, repeated small merges are less work [06:46] Kamion: agreed [06:46] Kamion: agreed [06:46] oh, ok [06:46] sabdfl: yes, but they're small, easy merges generally [06:46] Kamion++, keep the merges small [06:46] Kamion: agreed! [06:46] one big merge just becomes unintelligible and you end up breaking it down into repeated small merges anyway [06:46] agreed [06:46] ok [06:46] elmo: ping? [06:46] yeah [06:46] well, lots of merge work for the hoary team will make them expert endusers for hct when it comes :-) [06:47] elmo: is there any way you can make the warty morgue available ? [06:47] we should schedule a Hoary kickoff meeting [06:47] Keybuk: it already is on rookery [06:47] that will include reviewing the list of potential feature goals, deciding what we can actually do for hoary, and breaking down tasks [06:47] last updated a week or so ago [06:47] mdz: Thursday for sleep, Friday for meetings? :) [06:47] elmo: rookery doesn't have much on it though (like tla :p) [06:47] Kamion: thursday-sunday sleep, monday meeting? [06:48] and is rookery caneable? [06:48] mdz: no later that 14:00 UTC please :-) or my future wife will hunt you down :P [06:48] mdz++ [06:48] mdz: i figure you and the team will make the best call on merge strategy [06:49] please push the arch team to make arch suit YOU, not the other way around [06:49] adapt the problem to the solution :-) [06:49] sabdfl: yes, I have no problem with the strategy, only the fact that we don't have time to breathe between releasing warty and opening hoary [06:49] because opening hoary requires the huge merge effort [06:49] i think give your team the break [06:49] but we can take our time for it, right? [06:49] let's spend some time thinking about what we want in hoary [06:49] fabbione: that's 0700 here, which causes problems for MY home life :-) [06:49] the merge will go much faster if we have exciting goals for hoary [06:49] sabdfl: so, announce that hoary will remain more or less equal to Warty for a week or so? [06:50] i'm asolutely fine with that [06:50] ok, works for me [06:50] the warty team has put in a superhuman effort [06:50] they're chomping at the bit, but they'll survive for a week I think [06:50] I thought we sync all unmodified packages? [06:50] the last few days should have been a slight break since we were in such deep freeze [06:50] pitti: I don't think we should do that until we've reviewed the modified ones [06:50] otherwise we're likely to break deps all over the place [06:50] but everyone needs a proper release [06:50] mdz: we could employ the sarge testing scripts to check that deps don't break [06:50] mdz: that'll at least warn people hoary is a devel branch ;) [06:51] ok, so let's talk about the non-merge challenges for hoary [06:51] pitti: we're already running them in report-only mode [06:51] so much the better [06:51] pitti: http://people.ubuntulinux.org/~cjwatson/testing/ [06:51] I'd like to discuss the feature goals at the hoary kickoff meeting when everyone is around [06:51] Kamion: so would it be possible to sync all packages that don't conflict to anything? [06:51] mdz: daniels and I will be busy on the X-Man meeting from the 1st of Nov. [06:51] ok, i'm more interested in discussing how we organise it than what the goals are [06:51] mdz: for 2 full weeks [06:51] sabdfl: which aspects? [06:51] pitti: I suspect that's more trouble than it's worth; sync 'em, let the team sort out the breakage [06:52] well, some of them are similar to warty [06:52] Kamion: okay, agreed [06:52] debian sid + stability and security [06:52] gnome 2.x [06:52] x.org [06:52] Kamion: I don't feel that tooo much will break [06:52] but some are also whole chunks of brand new work [06:52] Kamion: (as long as we don't sync older versions) [06:52] and that's something we didnt do for warty [06:52] new apps and applets [06:52] bootsplash [06:52] we did some smaller chunks of new work for warty [06:53] but much of that was opportunistic [06:53] right [06:53] which is easier to manage [06:53] i prefer to work opportunitically, frankly [06:53] but for hoary we have set some goals [06:53] for hoary, I think we'll trim down the list of feature goals to a subset which is achievable for hoary [06:53] and build a small team for each [06:53] either way, i think we'll find that the goals will fall on the floor if we don't start early [06:53] assign each goal to someone [06:54] there are a number of community members who are interested in working on the feature goals [06:54] it'll be an interesting exercise in the first month or so discovering how much of what we've set ourselves is achievable; we haven't settled into time estimation yet really [06:54] or put a reasonable bounty on it [06:54] Kamion: agreed [06:54] it's that new aspect of things that i'm concerned about [06:54] we can do it [06:54] Kamion: as far as I'm concerned, everything on the page now is just brainstorming [06:54] we'll choose from that what we can actually do [06:54] but it's a new management thing [06:54] mdz: agreed, it's a wiki after all [06:54] much of it will fall to grumpy [06:55] estimating is very hard to do [06:55] i'd like to identify the things that are absolutely core and have our team do those [06:55]