=== warthylog [~warthylog@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-devel === Topic for #ubuntu-devel: Ubuntu development -- general discussion and support on #ubuntu | http://www.ubuntulinux.org/wiki/DeveloperResources | http://www.ubuntulinux.org/wiki/HoaryGoals === Topic (#ubuntu-devel): set by mdz at Fri Jan 7 18:46:16 2005 === justdave [~dave@66.227.241.236.gha.mi.chartermi.net] has joined #ubuntu-devel === sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-devel === ironwolf [~ironwolf@c-24-6-169-124.client.comcast.net] has joined #ubuntu-devel === ironwolf [~ironwolf@c-24-6-169-124.client.comcast.net] has joined #ubuntu-devel [09:53] sladen: we're using create_compressed_fs, which is actually advfs installed with a different name [09:53] morning guys [09:54] lamont: still around? [09:54] fabbione: morning [09:54] hey mdz [09:55] hey, who sang that song, "it'll all be better if we slept together"? [09:55] jdub: you are begging us to say you mom (part 2) [09:55] :-) [09:55] sorry.. couldn't resist [09:56] the name of my new band is Jeff's Mum [09:56] perl: I HATE YOU [09:56] dude, you so want to *be* my mum === boglot [~logbot@gw.workaround.org] has joined #ubuntu-devel [09:57] who is the owner of this boglot? [09:57] ~logbot@ [09:57] i don't mind it.. just curious === fabbione enables pkgstriptranslations on sparc buildd [09:59] 19:58 < Greves> i may have to try this so-called "ubuntu" [09:59] we're lying, it's actually called "smillysloffilyploff" [10:05] bah time to start sandpapering another room [10:05] later fellas [10:08] fabbione: have fun === Keybuk [scott@descent.netsplit.com] has joined #ubuntu-devel [10:13] Kamion: #5517, want to look? === ultrafunk [~pd@eth779.vic.adsl.internode.on.net] has joined #ubuntu-devel [10:36] mornin [10:37] hey dude [10:46] Kamion: around? [10:47] after all my wrangling, it's starting to look like the initrd doesn't have the new cdebconf [10:50] $1 says I need to specify your magic flag to use the daily d-i build [10:53] ah, found it, DI_TYPE [10:54] moin [10:55] LAMONT [10:55] http://king.warthogs.hbd.com/%7Ebuildd/livecd/livecd-current.cloop: [10:55] 09:53:50 ERROR 404: Not Found. === mvo_ [~egon@ip181.135.1511I-CUD12K-01.ish.de] has joined #ubuntu-devel [10:57] mdz: there's a livecd.cloop which is a symlink to 20050115/livecd.cloop [10:57] thom: can you make it point to something which actually exists? [10:58] fuckit, i'll try swapping in a new initrd on my iso [11:00] mdz: livecd-current.cloop should work now [11:00] as should livecd.cloop [11:02] thanks === arkuin [~arkuin@169.Red-217-127-113.pooles.rima-tde.net] has joined #ubuntu-devel [11:09] mdz: dude, tell me you're not still up? [11:09] oh, but I am [11:09] and I can almost taste it [11:16] do we use cfq by default in our kernels? [11:17] /boot/config-2.6.10-2-k7:CONFIG_IOSCHED_CFQ=y [11:18] yeah, but IIRC AS is the default [11:19] I'd check, but well, I don't have any non-custom kernels handy ;) [11:19] Using anticipatory io scheduler <-- from mine [11:19] ROCK ON [11:19] X autoconfiguration in the live CD [11:20] sweet! [11:20] woot [11:20] mdz: morning [11:20] Kamion: just in time [11:20] mdz: did you tweak CONF.sh? I was reading the backlog thinking "shit, forgot to change that over so that he wouldn't have to know [11:21] mdz: that main->universe symlink stuff was in order that warty CD builds remain possible [11:21] if somebody trimmed the broken symlinks, they broke that again :) [11:21] elmo: yeah, anticipatory is the default [11:21] Kamion: I think you'll have to only use --copy-links for pool [11:21] elmo: ah, ok [11:21] mdz: how scary is the livecd image-building foo? [11:21] I saw the anonftpsync change, will review/commit [11:21] Kamion: k, sorry for using your account like that [11:22] not a problem [11:22] jdub: there's the d-i building bit, the live image building bit, and the CD image building bit [11:22] elmo: and yeah, we probably ought to default to CFQ IMO [11:22] mdz: yeah, the live image [11:22] how come it needed my account in particular? was there stuff non-group-writable? [11:22] 'cos if so that's a bug I'd like to fix, so mdz can do this stuff [11:22] Kamion: I dunno, matt asked me to sort it, and I'm not in the group - i didn't want to use sudo in case I left broken permissions around [11:23] jdub: shouldn't be particularly scary, but then, no one has seen it except lamont ;-) [11:23] ok, no problem, I was away and he needed to get stuff done [11:23] mdz: ... ah. [11:24] mdz: I've set DI_TYPE to daily-installer, next daily build will use the new d-i build [11:24] Kamion: once casper 0.14 builds, a new set of CDs would be great [11:24] mdz: I rsynced a full live CD last night, and tried to rsync a new one today; it is not rsyncing usefully [11:25] Kamion: yeah, lamont discovered this as well [11:25] it's worth some investigation to find out why; intuitively it seems like we should be winning === opi [~emil@212.244.193.82] has joined #ubuntu-devel [11:26] morning [11:26] jdub: it was rad, I just booted up this hoary live CD straight into GNOME, and the desktop looks newer than the one on my workstation ;-) [11:26] fresh crack [11:27] mdz: SASS [11:28] when I booted it and did a apt-get update, some seconds later update-notifier showed some updates :) funny [11:28] mvo_: last night, mine showed 156 updates ;) [11:28] COW is going to *love* that ;) [11:29] hehe :) [11:29] any ideas why my upload of kdelibs to warty-security hasn't built over night? [11:29] thom: I saw something strange with readahead [11:29] on powerpc and/or amd64, it was segfaulting [11:30] and on _shutdown_ [11:30] no idea how [11:32] since it doesn't even seem to run at shutdown [11:33] mdz: apt 0.5.27.2's fine for sarge, isn't it? [11:33] er, [11:34] Kamion: yeah, that was absolutely its intended target [11:34] mdz: readahead does the sort-readahead-file-by-disk-order on shutdown, doesn't it? [11:34] jdub: the 'stop' section in /etc/init.d/readahead is empty [11:34] and it doesn't have a link in rc/06] .d, either [11:35] so clearly I'm insane [11:35] maybe it segfaulted at boot [11:35] mdz: the changelog message in 0.5.27.2 is not backed up by any actual diff to the source though [11:35] I see the stuff in 0.5.27.1, but I don't see the diff for this setlocale() stuff [11:35] really? hmm [11:35] the setlocale() stuff was icing [11:35] the 0.5.27.1 bits are what is needed for sarge [11:35] the diff is like a single page, it's very clear [11:35] ok [11:35] weird [11:35] go arch! [11:36] a single page? the translation updates are missing as well? [11:36] yes [11:36] wtf [11:36] mdz: there's K76fileordering [11:36] it has MaxArgs and MaxArgsBytes changes, configure/configure.in versioning, changelog, and nothing else [11:36] which does the sort on shutdown [11:37] it ought to move to the same init script [11:37] but... :-) [11:37] mdz: approved 0.5.27.2, if you could leave a fix for these issues 'til after the next dinstall I'd appreciate it [11:37] mdz: so both readahead *and* fileordering were segfaulting? [11:39] thom: just checked to confirm [11:39] K76fileordering segfaults [11:39] S39readahead does not [11:39] this was powerpc just now [11:39] right, hrm [11:40] my amd64 is still in a box, does it happen there too? [11:40] will check [11:40] thanks [11:42] no problem on amd64 [11:44] right, it could be the funky inode code. could you possibly file a bug and i'll talk to daniel silverstone (since it's his code) on monday [11:44] Kamion: so, I think I'm going to crash, but a new set of CDs based on daily-installer and containing casper-udeb 0.14 would be great [11:45] thom: how much of the faster-boot experimentation ended up making it into hoary? [11:45] we don't have caching of /dev nor the faster gdm artwork yet [11:46] most everything bar that should be in [11:46] gdm's artwork is slow? [11:46] mdz: ok, will build in twenty minutes or so when casper 0.14 shows up [11:46] elmo: resizing it is, yes [11:46] I'll queue an rsync for +1 hour or so [11:46] boggle [11:47] elmo: (it's a png, not svg, so gdm has to do some hard work) [11:47] mdz: ok [11:48] thom: I've added an entry to the HoaryGoals page for it; please create a wiki page to record which bits went in, and also the future stuff that we could do [11:48] so we can brag about it in the release notes, etc. [11:49] elmo: seems gdk-pixbuf is crap at scaling images [11:49] mdz: sure [11:49] thom: turns out an svg version of current gdm glow is 'hard' [11:49] elmo: could you update germinate on jackass? I added support for per-arch seed entries, would like to start using them soonish [11:49] thom: it's not just one gradient ;) [11:49] jdub: ah. [11:49] thom: so we'll probably have something new and special [11:50] thanks [11:50] night all [11:50] so we're racing to gdm, and then sitting around and twiddling our thumbs for an image to load :-) [11:50] not elmo special, proper special [11:50] night mdz [11:50] night dude [11:52] * Enable X autoconfiguration(!) [11:52] :-) [11:54] n8 mdz [11:55] elmo: will anastacia (or whichever one it is) be able to deal with germinate giving different answers for main-ness/universe-ness on different architectures? [11:55] (doesn't really matter if it can't and I guess overrides probably don't get to be per-arch like that, would just be interesting) [11:57] elmo: ping [12:01] mdz: g'night [12:02] also, either elilo needs to be removed from i386 or efibootmgr needs to be built on i386, one or the other [12:03] I think probably the latter given that elilo explicitly claims that it's useful on some i386 systems [12:04] hm, do Intel x86_64 systems use EFI? [12:04] and/or elilo ... [12:08] Kamion: nope, it almagates all the architectures, so if you're in main for one, you're in for all [12:08] fair enough [12:09] Kamion: I don't believe so---remember those machines are still coming up as 16-bit 286's [12:09] still, it will help out (a) CD images and (b) debootstrap, and you'll need to understand the new syntax [12:09] btw, the gdm image stuff couldn't you just resize permanently? most people don't change res often? combined with a mini-rainbow of common sizes being generated by cron after install or something? [12:09] sladen: so are i386en and apparently they can use EFI [12:09] Kamion: ok, will do when I get to the DC.. [12:10] sladen: it's a matter of boot firmware rather than 64-bitness AIUI [12:10] Kamion: I don't think they do use EFI no [12:10] HP DL3x0's G4's are em64_t, but AFAIK they have a normal bios, I'll ask taggart [12:10] hm, ok [12:10] is there an eaysy way to get laptop-mode not to mount partitions with noatime? [12:11] eh, in warty [12:11] http://adare.warthogs.hbd.com/%7Ebuildd/livecd/livecd-current.cloop: [12:11] 11:11:01 ERROR 404: Not Found. [12:11] can someone fix that? [12:11] amd64 and i386 were fine [12:11] how would it be fixed? [12:12] livecd-current.cloop-1024:65536 exists ... [12:12] Kamion: doing so now [12:12] azeem: you actually have a use for atime? on a laptop? [12:12] thom: by manually adding the CD ? [12:12] err, link? [12:12] elmo: it breaks mutt [12:12] if so, please be sure to tell lamont so he can actually fix it [12:12] azeem: boggle [12:12] elmo: yes [12:12] and yes [12:13] I think the -4096:4096 build failed and that was what was supposed to be linked to the plain name [12:13] Kamion: done [12:13] judging from terranova === arkuin [~arkuin@169.Red-217-127-113.pooles.rima-tde.net] has left #ubuntu-devel ["Abandonando"] [12:13] thom: thanks [12:13] same happened on amd64, fwiw === Kamion babysits [12:15] azeem: how does it break mutt? [12:16] bob2: mutt sometimes thinks there's new mail in mailboxes all the time [12:16] aha [12:16] "If you use mutt and experience this, you must disable the noatime remounting in the control script by setting DO_REMOUNT_NOATIME=0." [12:17] control script == /usr/sbin/laptop-mode [12:17] took me a while to figure that out :) [12:18] it sources /etc/default/laptop-mode, no? [12:19] mdz: new live CD up, seems sane [12:20] thom: I think the laptop-mode in Hoary/Debian is out of date; there's new ones I've seen that have about 20 options in the config file [12:20] sladen: quite likely [12:20] there are options in the Debian one, yes [12:21] Kamion: just Googling, I've found one press release mentioning plans to produce an amd64/efi system [12:21] thom: yeah, that works as well, thanks [12:21] np === thom rips the script out of the current kernel and tests [12:22] Kamion: but I suspect you can ignore it until that actually happens [12:33] kamion: can we close #1232? === seb128 [~seb128@ANancy-151-1-17-47.w83-194.abo.wanadoo.fr] has joined #ubuntu-devel [12:49] doko: just looking at the last couple of entries [12:55] sladen: "coloured prompt!" [12:55] sladen: "no, what do you want from the boot process?" [12:55] sladen: "optimised packages!" [12:56] jdub: "use flags!" [12:56] do users want a boot process that can be inserted nasally? [12:56] right into the main vein [12:58] we need a focus group. This is like extracting water from a stone [12:59] Kamion: have you done mini.iso installs? [01:01] jdub: which mini.iso? === Keybuk [~scott@syndicate.netsplit.com] has joined #ubuntu-devel [01:02] Kamion: warty's [01:02] jdub: no, *which* mini.iso? netboot, monolithic, ...? [01:03] mini.iso is just a way of putting the bits together, it doesn't specify which bits [01:03] people have done netboot installs, I haven't done one with final warty and I know there were a couple of bugs [01:04] I did a hoary netboot install yesterday, was basically ok [01:05] Kamion: oh, netboot i guess, couldn't see any others [01:06] probably depends on the arch [01:06] yeah, some dude did a warty install and it was incomplete [01:06] i386 [01:06] hm, how so? [01:06] I can give it a try later ... [01:06] didn't have ubuntu-desktop [01:06] and a bunch of other stuff [01:06] boggle [01:06] that's a function of the archive not the boot method [01:06] he may have done something wacky [01:07] was he trying to use the CD as a "mirror"? [01:07] no [01:07] full net install [01:07] ok [01:07] it was greves on #ubuntu if you want to ask for more details [01:07] I'll check it out next opportunity, then [01:07] I have to go in twenty minutes unfortunately, maybe later [01:07] thanks for passing that on [01:09] one thing that was strange with the net-install is that it only offered 'Great Britain' as a mirror option. Eventually figured that was because that's where the Canonical machines are. Is it worth adding an automatic skip in the case of only one choice? [01:15] sladen: theoretically, but I'd rather add more mirrors === ogra [~ogra@p508EA37D.dip.t-dialin.net] has joined #ubuntu-devel === two-face [~user@82.227.97.206] has joined #ubuntu-devel [01:21] Hi. === elmo [~james@82.211.81.249] has joined #ubuntu-devel [01:22] elmo, lamont: ping [01:22] oh, hi elmo [01:22] ? [01:22] elmo: so Riddell did a warty-security upload - do you still have to confirm it? [01:23] elmo: are security build logs ever published? [01:23] jdub: warty-security is setup to require confirmation from the security team yes [01:23] and no [01:23] cool, ta [01:23] (security team meaning pitti) [01:23] oh right [01:23] kdelibs got built fine, it's just sitting in queue/accepted [01:24] great [01:24] just answering Riddell's mail on u-d [01:24] elmo: why are the logs not published? [01:24] I've no idea how we want to handle warty-security, i.e. whether or not we require confirmation [01:24] Riddell: because security uploads are often embargoed [01:24] heh [01:24] what does embargoed mean? [01:24] err, handle warty-security +"for universe" [01:24] Riddell: answering your mail atm, with that precise phrase in it :) [01:25] Riddell: not to be released until a certain agreed date [01:27] elmo: ah, that makes sense [01:28] hmm, if I'm an approved ubuntu person does that mean I get on planet.ubuntu :) [01:30] Riddell: planet ubuntu will actually be for members soon, not just developers :) [01:39] is the website team aware of just how heavily ubuntulinux.org uses text-shadow? it gets quite annoying after a bit [01:39] Riddell: whereabouts? [01:41] Kamion: so, what's going to change when I update germinate? [01:42] elmo: with the current seeds, nothing [01:42] elmo: well, depends what your current version is I guess ;) [01:43] elmo: once seeds get updated, there'll be stuff like "elilo [ia64] ", so elilo will only be considered part of the base seed if you're germinating for ia64 [01:44] it won't be actively excluded on other architectures though - i.e. if something on i386 depends on elilo it'll still be included [01:44] just not seeded [01:44] ok, so that won't matter to me [01:44] hmm, which doesn't really help does it? [01:45] elmo: it helps with the CD images and my auto-debootstrap-updater - the only reason I want you to update is that the syntax for arch-specific seeds would be invalid in old germinate, I think [01:45] presumably the point of that was to not pull things in that can't be dep-satisfied? [01:45] ok [01:45] no, the point is not to have to pull stuff into base on one arch just because it's base in another [01:45] or similar [01:45] it doesn't matter for main/universe-ness, it matters for which bit of main/universe it's in [01:46] which bit of main, I mean [01:46] ok [01:46] jdub: all the headers http://muse.19inch.net/~jr/tmp/ubuntu-text-shadow.png it would be fine if it was toned down a bit [01:46] I thought it might be trying to address oo.o on ia64 for example [01:46] up to now I've had a system of overrides in debootstrap to cope, but that's a bit informal and really ought to die [01:47] ok, I'm up to colin.watson@canonical.com--2004/germinate--mainline--0--patch-29 [01:47] elmo: lamont mentioned that, but I'm not convinced that'll work; otoh we might be able to take it out of desktop on ia64, I think, which might help [01:47] it'll still be a build failure then but wouldn't block livefs builds [01:47] elmo: that's currnt [01:47] +e [01:48] kamion: it won't work [01:48] the archive can't do per-arch components for a given binary package [01:48] hmm, anasatacia got very upset [01:49] elmo: like I say, I don't care whether it's still in main or not [01:49] Riddell: holy crap [01:49] Kamion: yes, but to solve oo.o, you'd have to [01:49] elmo: if it's still in main that's fine, but we could make ubuntu-desktop stop depending on it [01:49] Riddell: i guess gecko doesn't implement that [01:50] well sure, but we'd still have uninstallable packages in main, which sucks [01:50] elmo: don't see why? if it's not depended-upon it won't be included [01:50] Riddell: that looks completely crap - can you file a bug? [01:50] oh, yeah, it's not a full solution, but it would stop being a live-CD blocker [01:50] sure [01:50] what broke anastacia? [01:50] anyway, this doesn't seem to have worked too well, linux-source is getting no love for !i386.. checking [01:50] huh, odd [01:51] uh? [01:51] the last change really ought to be a no-op with current seeds [01:51] kamion: I was quite out of date [01:52] I jumped 15 odd patches [01:53] hm, the multi-component/multi-dist stuff would be the prime candidate for something going wrong [01:53] it was all fine for me on cdimage though [01:54] elmo: anyway unfortunately I really have to go, late already; if it's a blocker feel free to revert and I'll sort it out later [01:54] do you care if Desktop goes screwy? [01:54] other than it's not a blocker [01:55] yeah, kinda [01:55] cdimage cares anyway [01:55] if you could mail me the germinate invocation you're using that would be goodd [01:55] cares before you get back? :) [01:55] mmm, probably not === HostingGeek [~HostingGe@200.48.233.220.exetel.com.au] has joined #ubuntu-devel [01:56] I'll be away most of the afternoon/early-evening though [01:56] ok then, i'll leave it, i'm not even sure how I'd revert.. I'll poke at it and mail you details including how to reproduce [01:56] baz update === two-face [~user@82.227.97.206] has joined #ubuntu-devel [01:56] yeah the problem is the last part [01:56] oh, don't have the output from the update around any more? [01:57] oh, yeah I do, blah.. too tired [01:57] or rather, I have a backup copy of the entire dir, I'll revert to that for now [01:57] anyhow ... gone === trulux_ghost [~lorenzo@67.Red-80-25-56.pooles.rima-tde.net] has joined #ubuntu-devel === doko [~doko___@dsl-084-057-096-239.arcor-ip.net] has joined #ubuntu-devel [02:15] jdub: what is the difference between a developer and a maintainer? [02:15] Riddell: developer covers committers, maintainers, etc. [02:15] Riddell: so it's just a more general term [02:15] jdub: you said planet would be for "maintainers not just developers" [02:16] Riddell: members not just developers [02:17] jdub: so who's allowed on the planet at the moment (not because I want on it, I'm trying to understand the terminology) [02:18] Riddell: atm, it's an unupdatable blob ;) [02:18] Riddell: i'm fixing things up atm, so that when it will be maintained from an arch branch [02:18] at which point, it will be for any member [02:19] ok, richt [02:19] i'll put your blog on straight away :) === HostingGeek [~HostingGe@200.48.233.220.exetel.com.au] has joined #ubuntu-devel [02:19] do you have an ubuntu or tech category? [02:20] elmo: btw, could you send the index.html.tmpl from the planet install? i didn't realise you just sent the config last time :-) [02:20] jdub: you need to put a flowchart diagram up that goes member -> committer -> maintainer or something [02:20] yeah [02:20] jdub: http://www.kdedevelopers.org/blog/feed/57 [02:21] hrm [02:21] jdub: I thought I tarred up config-ubuntu for oyu.. but ok [02:21] drupal does such poo rss ;) [02:21] at least it has dates now [02:21] elmo: hrm, wait a sec [02:22] i'll check personal mail [02:22] nup [02:22] jdub: sent [02:22] thanks [02:22] now to deplone it === HostingGeek [~HostingGe@200.48.233.220.exetel.com.au] has joined #ubuntu-devel [02:24] doko: I just passed you a pyopengl b-d bug [02:24] doko: do you know off hand whath appened to python2.3-dictdlib? dict-moby-thesaurus still b-d's on it [02:25] python2.3-dictdlib is nowpython2.4-dictdlib, will update dict-moby-thesaurus [02:25] doko: ah, well it must have FTBFS [02:26] 'cos python2.4-dictdlib doesn't exist yet in the archive :) [02:26] could you chase that down too? [02:26] pyopengl b-d, I should pass that on to daniels, see http://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=xvfb [02:27] elmo: ok, btw, what any news about the powerpc64 chroot on davis? [02:28] jdub: should I turn up at at the next technical board meeting and ask to be made a maintainer so I can work on kde packages in main (and does haggai need to do the same)? [02:28] doko: err, what would I build it with? I thought you were just asking if I would do it, not for me to do it [02:28] Riddell: haggai has been approved already [02:28] doko: can you make the pyopengl bug say 'blocks on ' (whatever is for our BTS) [02:28] elmo: will do [02:28] Riddell: you should put your name on the maintainer proposals page, and CC and TB agenda [02:29] doko: [ppc64] will a powerpc chroot do to start with? [02:29] jdub: ok. why the CC agenda? [02:29] elmo: chroot: yes, then install the things I wrote you in an email. [02:29] Riddell: CC and TB approve maintainers [02:29] doko: I don't seem to be getting email from you [02:30] doko: where/when/to-who did you send it, and what's the message ID? [02:30] Riddell: gotta get the love from everyone ;) [02:31] elmo, jdub: so CC approves members, TB approves committers and both CC and TB approve maintainers? [02:31] sladen: the flow is fairly obvious, it's the details that arn't clear [02:32] Riddell, wasnt there a propsal from mark that you get both stati in one step to speed up the MOTU process === two-face [~user@82.227.97.206] has left #ubuntu-devel ["ERC] [02:32] Riddell, thats how i understood it at least [02:32] Riddell: the bar for being a universe maintainer is quite a bit lower than for being a main maintainer [02:32] elmo: sent again. you did get my email about the mailman sync from unstable? [02:33] Riddell: haggai was approved quickly because he's an experienced debian maintainer, and has a saucy and convincing british accent [02:33] hehe === trulux_ghost is now known as trulux [02:35] ogra: do you have a blog? [02:35] yup [02:35] ogra: with an ubuntu or tech category in it? [02:35] nope, do i need one ? [02:35] not necessarily [02:35]