[00:02] jdong_: With LP project<->team affiliations which are coming soon, comments by real people (ie. people connected to the project somehow) will be able to be shown as more important. [00:10] could somebody explain the reasoning why debtags and ept-cache are in the ubuntu repository but not integrated in ubuntu at all? [00:12] perhaps nobody has had time/interest to integrate it yet [00:12] yeah, i started but i got bored [00:12] i might write it in intercal [00:12] if 1 person would care to promote it in a suitable way there might be enough peope on #ubuntu-motu who contribute? [00:13] ideas are cheap. developers less so [00:14] enrico did all the word it needs from the developer side [00:14] i.e. feel free to make proposals, e.g. use brainstorm, but don't anticipate a flood of people to implement your ideas before anything else [00:15] before i waste my time the next weeks i am asking for the reasons why its potential use ignored by the core developers [00:15] use/is [00:15] I didn't even know ept-cache existed [00:15] I know about debtags but I'm not sure if we've got anything that would really make use of them [00:16] its one of the smartest ideas on how to handle thousands of package descriptions in such a way that users find what they are looking for - since binary packages were invented [00:17] because traditional categories for packages are fine for Ubuntu's CD 1 [00:17] but not for 20,000 packages [00:17] I think most people see the value in debtags (or at least the concept) [00:18] why the promotion without the debian community did not work is kind of logical. anything new is a danger to conservative people. but i dont get why the concept is completely ignored by ubuntu [00:18] without/within [00:19] I don't think it's completely ignored [00:19] just that nobody's had time to do anything interesting with it; as you say, it isn't really needed by the default installation so hasn't reached top-priority kind of status [00:19] I think we'd welcome somebody working on it properly [00:19] there's been no explicit "debtags sucks, let's ignore it" kind of decision AFAIK [00:19] Keybuk: will the new udev also declare a Breaks: on the old lvm2 (eventually)? [00:19] i am not even close to become an ubuntu developer. so its not the right task for me [00:20] Keybuk: also, are the changes to libs-cleanup.patch and drop-realtime.patch in lvm2 intentional? [00:20] it requires a person with enough influence to poke the right people so the necessary promotion starts. the volunteers who do the work will come to you automatically [00:20] I don't think it works like that at all [00:21] lots of things are bottom-up not top-down, in my experience [00:21] and volunteers CERTAINLY do not automatically materialise no matter what you do [00:21] slangasek: I had to refresh them, quilt told me to otherwise it would kill kittens [00:21] promotion is the key [00:22] Keybuk: ah; the refreshed patches look rather... different from the previous revision [00:22] Breaks - yes, it should probably have some of those [00:22] slangasek: really? I didn't look at them [00:22] anyone can do promotion [00:22] Keybuk: both patches now patch less than they did before :) [00:22] not sure why that would've happened with a refresh [00:22] slangasek: quilt is evil [00:22] although I would suggest a more constructive attitude than "before i waste my time the next weeks i am asking for the reasons why its potential use ignored by the core developers" / "but i dont get why the concept is completely ignored by ubuntu" [00:22] not IME [00:23] thats not true. when the right people post a comment on their blog it creates huge waves of attention [00:23] when i dedicate 5 servers it gains me nothing [00:23] consider how the right people became the right people [00:23] they did the work [00:23] anyway, this isn't my area of interest, I'm just advising; you can take it or leave it [00:24] fantastic for those who did the right things ages ago. maybe i become a hacker in my next life [00:25] fruchtix: Whining about not being influential and not actually doing stuff isn't likely to help you reach your stated goal. [00:25] so far your general tone has succeeded in persuading me to go off and do something else more productive. :-( [00:25] its just sad when you see for what rubbish some celebrities use their blog and a brilliant idea dies slowly because it gets no attention [00:26] fruchtix: Around here doing stuff tends to get you more attention than whining or blogging. [00:26] slangasek: looks like if you pop -a after building, quilt screws up - reuploaded with the original 2 patches [00:26] dude, this is not about writing code. nobody needs your code in this case. [00:27] the code already exists. and a lot of data exists too [00:27] but you see what reactions i get here on this channel? what do you think what reactions i get when i start "doing work", huh? [00:28] fruchtix: code does need to be written to integrate debtags into useful apps like gnome-app-install and synaptic [00:28] huh? the reaction you got from me was entirely due to your horrifically negative tone [00:28] so thats why i takes a person with a reputation to pick it up [00:28] fruchtix: or at least someone with a less-whiny tone [00:28] the people with reputations have an awfully large amount of work that they have to do already [00:28] my tone? you feel my tone by text only? [00:28] thats amazing [00:28] I'm entirely happy to support debtags work in any way I can usefully do so, but it isn't my area of expertise so I'd rather do something I'm expert in [00:28] i love this cyberspace thing [00:28] fruchtix: No, that's why it takes someone whoe doesn't have a horrifically negative tone to take it up. This may be in your grasp. [00:29] fruchtix: yes, you're full of biting sarcasm, thatd be tone. [00:29] text is the way in which you're communicating, so you need to be careful about how things come across; I'm afraid that's life [00:29] you're such an optimist, colin [00:29] Particularly when you're trying to convince other people they want to work on something that interests you more than the. [00:29] charisma might help [00:30] maybe the situation creates the sarcasm and the feeling of anger for people who believe the gained all the fame for themselves instead of using it for the benefit of the project and the users [00:30] or maybe you just want free developers to work on your pet projects [00:30] who can say, in this crazy world of ours [00:30] thats actually the sharing philosophy, even when your paycheck from canonical is sexy [00:30] hey uh you realise a big chunk of this channel gets no paycheck for FOSS? [00:30] fruchtix: Most of the people talking to you don't work for Canonical. [00:31] after all, its the code of upstream in most cases that makes you happy and proud [00:31] so how about you check your own tone when you talk to me? [00:31] fruchtix: who says people here dont work with upstream? [00:32] and how about this code of conduct. did i sign it yet or did you sign it? [00:32] fruchtix: So far everyone here has been trying to give you helpful advice. [00:32] oh yeah? [00:32] like I say, I'm entirely happy to support debtags work in any way I can [00:32] Yes. [00:32] by judging my person? [00:32] by complaining about my "tone"? [00:32] ScottK, i haven't, i've been gently dripping sarcasm as i usually do [00:32] by telling you that if you want to get something working, then you should go make it work [00:32] OK. Except him [00:32] You've commented about your perceived inability to influence people; in that context I think comments about your tone are entirely reasonable. [00:32] directhex: no more than fruchtix has :P === nhandler_ is now known as nhandler [00:33] dude, if you seriously believe in this rubbish "if you want something done then do it yourself" then why the heck are we talking about a community? [00:33] Because that's pretty much a direct input into your ability to influence people. [00:33] fruchtix: what do you think a meritocracy is? [00:33] "community" doesn't mean "people doing stuff i want when i want it" [00:33] of *course* if you want something done right you do it yourself [00:33] fruchtix: If you seriously believe popping into an IRC channel and saying "everyone do this" works, you have no idea how the world works. [00:33] that's how its been since humans evolved/were created/came from neptune [00:33] heard about co-operation? [00:33] heard about "i am dedicated to my work"? [00:34] fruchtix: Yes. You? [00:34] heard about "i care for ubuntu and i dont just hang around here because of my cool vhost"? [00:34] Keybuk: ta [00:34] * cjwatson looks at the upstream release he's preparing over --> there at 00:34 local time [00:34] because I care about my work [00:34] perfect [00:34] okay, now the unexpected reality blast: if you want to see "something" done with debtags, why not make a concrete proposal which people CAN look at and comment on and cooperate on? [00:34] don't preach at me [00:34] cjwatson: when did i address you personally? [00:34] rather than a handwavey "DO MY STUFF!" [00:34] quite a bit, as it happens [00:35] aha [00:35] you have a lose concept and a chip on your shoulder. sort BOTH to get results [00:35] I gave you useful advice, you got all offended [00:35] cjwatson: and the person who pays your salery (your boss) tells you straight into your face "lamer, listen. if i need something done i do it myself. go on vacation"? [00:35] you don't neccessarily need to write buckets of perl - just come up with a real proposal beyond "you guys should drop what you're doing, i command it so" [00:35] fruchtix: no, in that case you're being PAID [00:36] i'm spent, sarcasm mode back on. [00:36] fruchtix: unless you are intending to offer a cash bounty to the person who does what you want, don't be so bossy, please [00:36] so all of a sudden there is a difference between a paid job _for_ the community and a volunteer job _for_ the community? [00:36] so all of a sudden we all realize that we only move our arse for other people when the paycheck is sexy? [00:36] look, hackers'll hack on whatever they want whenever they want [00:36] is that the free software spirit now? [00:36] they don't need to take orders from anyway [00:36] I believe the only person in this conversation who is paid by Canonical is me [00:36] *anyone [00:36] hurrah for entitlement culture [00:36] kids these days [00:36] * ajmitch scrolls up to see what started this rant [00:37] well, thanks for the interview then. have a nice day [00:37] ajmitch, "USE TEH DEBTAGS NAO!" [00:37] so the paycheck rant is rather missing the point [00:37] directhex: ah ok [00:37] yeah... [00:37] and btw - thanks for trolling me, too [00:37] (and in any case I was involved with free software long before I was paid to do so) [00:37] you know, i think i WILL do "something" with debtags [00:37] the point is, people will work on what they'll work on when they want to work on it, and if they dont want to work on what you want to work on, then you can do it yourself or find someone else to do it for you [00:37] its the cheap tricks of a young journalist. but it seems to work [00:37] And oddly enough I've spent most of my FOSS time today working on stuff that other people asked me to do. [00:37] a bonsi buddy clone, written in visual basic.net, which picks a package at random and tells you its tags [00:37] sounds helpful! [00:38] directhex: you're a bad man [00:38] directhex: that's scary [00:38] ajmitch, by design! [00:38] just because person A finds something interesting doesnt mean person B will [00:39] directhex: paid by canonical or do you troll for hobby right now? [00:39] or that B will have the skills even if they find it interesting [00:39] fruchtix, for the love of it [00:39] directhex: cool [00:39] i am sure you have your fans [00:39] oh, i do [00:39] fruchtix: did ya miss the part where cjwatson is the only paid one here? [00:40] * directhex fires up monodevelop, looks for some pictures of a purple gorilla [00:40] And I'm pretty sure he's not on the clock at gone midnight local for him. [00:40] definitely not [00:40] O_O [00:40] how many hours a day do you spend on this colin? [00:40] ScottK, don't most hackers only start being productive at 10pm? or is that just me? [00:40] ... some [00:40] Depends. [00:40] doing a man-db upstream release at the moment, taking a bit longer than I'd hoped [00:41] so canonical gave up control over one of the most important ubuntu channels here on freenode. also interesting news and good to know [00:41] what? [00:41] but yeah, launchpad is ready [00:41] canonical never had control of this channel... [00:41] i can see the strategy [00:41] not really [00:41] fruchtix: Ubuntu has always been a community led distro. [00:41] directhex: I for one would appreciate it if you didn't taunt people quite so mercilessly when they're already clearly aggravated... [00:41] fruchtix: is there a specifc thing you want done with debtags? specifically or ept-cache or is there more? [00:42] canonical's probably got more tech support people than developers, i'd guess [00:42] LaserJock: how's the moodle merge going? :) [00:42] #ubuntu-devel was open to non-Canonical contribution from the start, and had significant community involvement from the start (I was there, so I can be pretty sure about this) [00:42] ScottK: that sounds quite romantic, but i think without Mark and his investment into Canonical ... most of you guys would suck on the nibble of Debian or other distros [00:42] slangasek: getting there. lots and lots of CVE reviews [00:42] slangasek: I'm waiting though on a MIR for smarty [00:42] fruchtix, most of the people you're arguing with are long-standing Debian Developers [00:42] fruchtix: Of course the Canonical investment is critical, but there is a very substantial community involvedm. [00:42] fruchtix: you mean if ubuntu didn't exist, period? [00:43] slangasek, is that a "shut up" or a "stop working on Debtag Buddy"? [00:43] certainly I've been a Debian developer since 2001 [00:43] ... involvement. [00:43] LaserJock: uh? there's an MIR of a template engine required to fix a moodle security bug? [00:43] directhex: maybe thats why i am in the middle of my schizophrenia again. its the debian personality aspects [00:43] directhex: 'yes'? [00:43] cjwatson: what would you say? more devs or 1-800-help-4-me people? [00:43] slangasek: not required, but I'm merging a new upstream release in [00:43] slangasek: the release fixes loads of bugs and security vulernabilites [00:43] LaserJock: so I'm thinking that's not going to all be ready between now and Thursday and I should probably bump the milestone? [00:43] slangasek: the template engine is a split out lib [00:43] * ScottK decides to go pick up his daughter from the mall since that's more productive than this discussion. [00:44] ScottK: the computer-y one? i say "hi" [00:44] maco: more developers [00:44] if you would be able to isolate the technical aspects of the work of a debian person, then well, then we would not need Mark and Canonical to create ubuntu as a "helper" for the broken social aspects, right? [00:44] LaserJock: oh, it's code that's already in the existing moodle package? that doesn't require a full MIR, just a bug saying that [00:44] fruchtix: the social aspects arent the problem [00:44] slangasek: well, kees asked for it. It's all there [00:44] fruchtix: its the ease of use stuff [00:44] slangasek: but then he went on vacation so I'm not sure where it stands [00:45] fruchtix: i mean, yeah, there's the coc stuff....but really....the 6mo release cycle and the aim for simplicity is what brings people to ubuntu, i think [00:45] its the reputation for being fairly easy to use [00:45] ....that's why my mom uses it [00:45] maco: you dont know much about debian then. i am observing the situation since 8-9 years. and its 95% home made ... and social aspects [00:45] LaserJock: "if a new source package contains only code which is already in main, it may not need a full report. Submitting a bug with an explanation is sufficient." https://wiki.ubuntu.com/MainInclusionProcess - so if that's the case for smarty, upload and we'll fix up the overrides... [00:45] fruchtix: what do you mean 95% home made? [00:46] maco: problems made by debian developers for other debian developers [00:46] he means that he's Patrick Frank [00:46] social problems were one reason we founded Ubuntu rather than simply funding work on Debian, but far from the only one [00:46] * maco is very confused [00:46] slangasek: how is helix? are you still doing her? [00:46] slangasek: it's bug #327367 [00:46] Launchpad bug 327367 in smarty "MIR: please promote smarty to Main" [Undecided,New] https://launchpad.net/bugs/327367 [00:46] slangasek: that would explain a lot [00:47] sorry, should've gone with my gut 20 minutes ago [00:47] that took longer than required [00:47] O_o [00:47] I hadn't twigged to the debtags connection [00:48] what's the connection? [00:48] google [00:48] has Patrick expressed interest in debtags before? [00:48] yes [00:49] didn't he once threaten OFTC with legal action? [00:49] freenode, yes; OFTC, I don't know [00:50] preemptive dispatch from #ubuntu-motu ? [00:52] no, he's just ranting away in -server now [00:55] slangasek: the moodle changelog has 26 security fixes and 20 debian bugs fixed since our current version [00:56] slangasek: I was guessing that would get me a FF exception [00:57] in Debian we've got a buildable package that has incorporated almost all of Ubuntu's diff [00:57] so I just need to test to make sure it's not totally broken, get FFe approval, and then I can upload [00:58] my apologies to this channel for being drawn in too far by a troll [00:58] *sigh* you make one little change to the task_struct, and you have to rebuild almost the entire kernel :-/ [00:58] [00:59] Keybuk: If you are using the build system used by the Ubuntu kernel, that happens even if you make a config tweak unfortunately. That can probably be fixed up though. [00:59] cjwatson, maybe you just see the good in people. i called you an optimist half an hour ago ;) [01:01] I'm only sorry that this means he'll probably infect our mailing lists for a while [01:01] mm, really? [01:01] its a moderated list... [01:01] ubuntu-devel is; ubuntu-devel-discuss and others would require more effort to keep clean [01:01] yeah [01:01] it might also ruin people's desire a little to actual do something with debtags [01:02] right, that concerns me more [01:02] I do think it's a useful system [01:02] LaserJock: just because itd make the jerk happy? [01:02] not necessarily [01:02] but getting involved in a "war" with him every time you try to do something would be pretty demotivating [01:02] as well as people just getting tired of the subject [01:03] And the author [01:03] the author? [01:03] i couldn't spot the name of the real author. google failed to find much from what /whois revealed [01:03] Enrico Zini [01:03] directhex: if you're referring to our guest, the name is Patrick Frank [01:04] (who as far as I can tell is nothing to do with fruchtix, and is generally a very laid-back nice guy) [01:04] a google on that name and debian or ubuntu is... elucidating [01:07] slangasek: so do you think I can get a FFe by Thursday? [01:07] hm. flashes of obnoxiousness [01:08] LaserJock: "by 6pm" is also "by Thursday", so sure [01:08] i think dragging erinn into the above little troll session was pushing it rather hard though. i do wonder about people like that [01:08] then i mentally file them in the same bin as boycottnovell contributors, and sleep easy! \o/ [01:09] directhex: stick around long enough and he'll give you plenty more fodder for wondering [01:09] slangasek, we'll see where my NM application takes me! [01:11] directhex: the condensed backstory there, btw, is that he had a habit of joining #debian-women, each time with a different alias, despite being told his presence was unwelcome and inappropriate; and as an op there I both a) got good at identifying him, b) pissed him off to no end [01:11] oh, he's been here too? [01:11] yes [01:11] And -server [01:11] * maco snorts [01:11] Hobbsee: that sentence told you who was being talked about? [01:12] the buzzing was probably enough to make that clear :) [01:12] oh, paddy. right [01:13] He is one of those people that wanted debian-women to stop existing, isn't he? [01:13] * StevenK tries to remember [01:13] StevenK, highly likely, given the above descriptions [01:15] james_w is warning #ubuntu-women ops [01:16] A k-line sounds appealing right about now [01:16] StevenK: well, that's the aim. [01:17] or at least, likely to be the end result [01:17] I was going to say, surely the aim is to keep the network productive ... [01:17] Hah [01:18] A K-line would surely just get him more wound up, anyway; we already know he morphs [01:19] he's trolling #ubuntu-ops now, they can deal with it [01:20] he needs a hobby [01:20] well, a DIFFERENT hobby [01:20] StevenK: not as such; he just wanted debian-women to not have ops that would kick him from the channel [01:21] Pity that he has moved on from annoying Debian to us. :-/ [01:21] StevenK, gotta go where the slangasek is! [01:21] slangasek: all the better to troll them with? [01:21] they all come our way eventually [01:22] StevenK: he goes back and forth on which community to harrangue [01:23] it'll be Sven Luther before too long [01:23] Friendly [01:23] can we send him to...i dont know...whatever community has Theodore de Raadst? that could be amusing. [01:23] (cant spell that name, you figure it out) [01:23] maco, openbsd [01:24] directhex: the point was theodore, not the community ;) i hear he likes to argue on the internet [01:24] maco, newsgroups, probably. these days, newsgroups are where you go if 4chan is too civilized [01:25] maco, it's where you can find some of my critics, for one thing ^_^ [01:27] 4chan? Civilized? [01:27] * StevenK chokes [01:27] StevenK: ok so we're laughing at the same thing [01:27] I think directhex was aiming for irony [01:27] Duh :-P [01:27] Keybuk, too sleepy for sarcasm [01:27] Keybuk, irony will have to do [01:28] * Hobbsee suggests the channel go back on topic? ;) [01:28] oh right [01:28] development [01:28] Hobbsee: that'd just be boring [01:28] sigh, I meant to do gparted 0.4.2 today [01:29] I guess that will have to be tomorrow, I'd rather not do it while half-asleep [01:29] cjwatson: I meant to do lots of things today :-/ [01:29] too much lvm inotify madness [01:29] did you make any significant progress by the end of the day? [01:29] Hobbsee, i've hit 90% of my jaunty goals already, though \o/ [01:29] there were a few too many trees for me to see the forest [01:29] directhex: \o/ [01:29] cjwatson: I believe I have fixed the bug [01:30] I'll be trying to confirm that as soon as amd64 binaries are available [01:30] yeah, the buildds are on a go-slow today [01:30] is openjdk still building on arm? [01:30] Yes [01:30] And koffice, and kdebindings [01:31] Started on 2009-02-18 [01:31] and that really isn't a dead build? [01:31] LaserJock: I don't see that a FFe request was actually filed? [01:32] sbuild will kill a build that spins for 150 minutes with no output, IIRC [01:32] it's ARM [01:32] it has to save up for compilation in installments [01:34] wait, was that a mortgage joke? [01:34] no, more of a "collect each monthly part and they build up to this nice collectable executable" [01:35] "pay by direct debit and get this attractive binder FREE" [01:35] oh, so it's not about adjustable-rate mortgages [01:35] no, are adjustable-rate mortgages especially on your mind right now? [01:35] i guess that means i no longer have that excuse to not use lvm (excuse being: it usually breaks badly during devel time) and should download a daily cd to install and test with [01:35] Keybuk: not particularly, I was just grasping at straws trying to parse "compilation in installments" :) [01:36] I wish I had a variable rate mortgage [01:36] mine's fixed [01:36] and now the interest rate is practically non-existant [01:36] Below inflation, for some people. [01:36] we have one on the condo that we haven't been able to sell, and it keeps costing us less and less every quarter [01:36] Keybuk: refinance? :) [01:36] slangasek: still in the fixed period [01:36] so should I be rebuilding d-i against the new udev once it builds everywhere [01:37] ? [01:37] there's penalty clauses for early closure [01:37] have only a year or so left [01:37] and will an email go out when there's a daily cd available with a hopefully-working udev/lvm combination? [01:37] Keybuk: sure; I assume you've looked at the math then, to determine whether it's worth paying the penalty [01:37] slangasek: yeah, marginally not worth it [01:37] I suspect that without a d-i rebuild the alternate install CD will have slightly hosed LVM [01:38] it wasn't really that high interest rate to begin with [01:38] maco: not explicitly; but if the bugs are still closed tomorrow, tomorrow's CD will have the fix [01:38] slangasek: Still can't sell the condo? :-/ [01:38] StevenK: nah - bearing in mind that it's still winter here, so the market is only just starting to pick up [01:39] slangasek: The housing market hibernates for the winter? [01:39] StevenK: yes [01:39] Odd [01:39] nobody wants to buy a house when they don't have enough light to look at it first [01:40] hurrah, my new kernel built \o/ [01:40] now to remember what I was going to do with it [01:40] make kernel panckaes [01:40] pancakes [01:40] slangasek: ok. / keeps filling so it seems i need to switch to lvm so i'll be able to resize it next time i fill it. [01:41] :) [01:41] slangasek: it is Tuesday here ;) [01:41] Tuesday is kernel pancake day? [01:41] maco: You have things like /usr and /var seperated? [01:41] Tuesday is Pancake Day [01:43] slangasek: http://en.wikipedia.org/wiki/Pancake_Day [01:43] Keybuk: wow, a cultural reference that totally eluded me :) [01:43] slangasek: I'm not entirely sure you knew about that [01:43] if not, that was an amazing piece of annual timing for a pancake joke [01:43] apparently so :) [01:49] slangasek: no, I didn't file one as I don't have the packages finalized [01:49] *package [01:50] LaserJock: oh. I guess that means I can't give you an FFe by 6 today [01:52] slangasek: why did you want it by 6 today? [01:52] LaserJock: because after 6, I'm not available to approve it until tomorrow :) [01:53] slangasek: 6 pacific time or central? [01:53] pacific [01:53] so 7 min [01:53] :) [01:53] hmmm, that's rather tight [01:53] as opposed to -113 min [01:53] LaserJock: don't sweat it [01:54] well, let me work on it and get it done now [01:54] if I get it in 7min fine [01:54] if not tomorrow will due [01:54] :-) [01:54] I need to get this bugger out the door [01:55] I've already spent quite a few hours digging through CVEs and inline patches [01:58] hmm [01:58] today's daily live is unwell [01:58] known ubiquity bug, milestoned [01:59] the fork-bombing nautilus? [01:59] install python-numpy by hand before invoking ubiquity if you want to use it [01:59] oh, no, that's not what other people reported with it :) [01:59] Nautilus now fork-bombs the system? [01:59] seems to [02:00] I get about 1,000 little "Starting File Manager" things along the bottom [02:00] Ow! [02:00] StevenK: it was mentioned in reference to update-manager mem usage, and didn't want to be outdone [02:01] Bwaha [02:02] Keybuk: can you spot any obvious reason why the fix in bug #291752 was wrong? [02:02] Launchpad bug 291752 in cryptsetup "[regression] cryptsetup does not work on raw encrypted drives" [Critical,Confirmed] https://launchpad.net/bugs/291752 [02:03] Keybuk: checking for the device node, but not checking that vol_id returns something meaningful for the device, caused a regression for at least one user who says his crypted rootfs doesn't mount right anymore in intrepid [02:04] StevenK: no, i dont. just / and /home. i have to apt-get clean after each round of updates because i have gnome and kde installed and that leaves only 1gb for updates...which it is very easy to hit [02:04] slangasek: the vol_id check in the initramfs? [02:04] that's probably important [02:04] hmm, evand fixed that ubiquity bug in bzr but didn't upload [02:04] guess I'd better do that now so that it makes the next daily build [02:05] Keybuk: I gather that it is important, but can you explain to me why checking for the device node isn't sufficient? There are two opposite use cases here, and while the short term fix is probably to revert to the intrepid-release handling, I'd like to get a handle on a more complete fix [02:05] cjwatson: I saw the commit message that added -numpy, and then he fixed the code, does it still depend on numpy? [02:05] device node can exist, but have no filesystem inside it, because other things need to do magic [02:05] e.g. md and devmapper devices [02:06] the device node shows up when you create the name [02:06] before you load the table [02:06] Keybuk: aha [02:06] that makes perfect sense then [02:06] StevenK: haven't checked, but will [02:07] so you need to check [02:07] 1. does the device node exist? [ -e ] [02:07] 2. has udev finished with it? udevadm settle [02:07] 3. has it got a filesystem inside it? vol_id [02:08] yep [02:08] thanks, that tells me what I need to know [02:09] hmm, u6y hangs at the partitioning stage now [02:10] StevenK: dependency is all gone [02:10] cjwatson: Yay [02:10] (r3057) [02:15] cjwatson: The changelog still mentions adding the dependancy [02:18] easily fixed [02:26] * Keybuk decides to go to bed [02:26] I'm now too tired to test this, and it's only 38% through installing [02:26] gn Keybuk :) [02:28] Keybuk: I *think* that might be fixed by ubiquity in bzr [02:28] I guess we'll find out tomorrow [02:41] You know, evolution is stupendously bad at handling my imap folder with ~200K emails in it. How would one debug "spends 30minutes thrashing the disc when updating message list"? [02:48] arthur-: Probably worthwhile archiving some of that mail. :) I am finding even mutt is starting to chug with big Maildir boxes, granted this is over a network, but still. [02:48] RAOF: ^^ [02:48] arthur-: sorry [02:48] Heh. [02:49] I'm guessing it's something to do with evolution's new sqlite-backed folder thingies; it used to handle this pretty well. [02:55] RAOF: grab evo trunk [02:55] RAOF: that was bugfixed [02:55] lifeless: Sweet. And will be in Jaunty release, because trunk is going to become 2.26? [02:57] Hey, is there anything documented anywhere about exactly how and when the cessation of support for Dapper is going to be done? [02:57] Like I know the LTS gets 3 years of support on the desktop [02:57] So that means sometime in June it runs out [02:57] So what will still get support (i.e. server-wise) after that date? [02:58] I think that's a good question. [02:58] RAOF: one hopes, don't know for sure. [02:59] My personal theory is that anything that needs X is toast. [02:59] Is it just me, or is there no logout/restart menu options in the GNOME system menu any more? [02:59] * TheMuso keeps doing keystrokes to get there, always forgetting they are not there, and ending up in about Ubuntu [02:59] I see that to, although was blaming gdm-new for it. [02:59] Caesar: server is apparently 5 years as per this page: https://wiki.ubuntu.com/Releases [03:00] RAOF: No, its either a gnome-panel, or gnome-session bug I think. [03:00] firefly2442: yeah, but what constitutes "server"? [03:00] More likely gnome-panel [03:00] Caesar: ahh I see the issue [03:01] And when in June 2009 does desktop EOL? [03:01] June 1 or June 30? [03:01] heh, it's just a bug in the X "server" ;) [03:01] These sort of things aren't clearly spelled out (at least not on that page) [03:03] Nick Barcet was working on that, specifically on creating full lists of packages [03:03] I don't know where he put the output though [03:03] nijaba: ^- [03:05] I'm also curious if the packages will be removed (as is normally done for an obsolete release) or left on the mirrors. [03:05] TheMuso: FUSA [03:06] maco: FUSA? [03:06] TheMuso: no more shut down in the menu because it was duplicating FUSA [03:06] I don't think we're likely to be able to remove them in any sane way [03:06] perhaps unfortunately, but there it is [03:07] maco: But what is FUSA? [03:07] removing them would require regenerating all the index files and other such invasive things [03:07] YOu say they duplicated FUSA, so I am confused. [03:07] Oh!!! Fast switch user applet [03:09] there ya go [03:09] grrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 [03:09] oops sorry [03:09] was called away for "this thing in the fridge that's been here for weeks..what is it?" [03:09] * TheMuso goes to another window [03:11] (it was vegetable sushi) [03:11] * TheMuso digs for a gconf key. [03:18] I have a general development question, how does Ubuntu deal with people contributing who live all over the world? It must be difficult for people who are able to contribute code but lack the communication skills because of a language difference to "sync" into the project and where it's headed. [03:18] people do need to have some English ability, in practical terms (if nothing else they need to be able to read everyone else's code, including comments) [03:19] and we do need to be fairly tolerant of language barriers, although yes it can be challenging [03:19] local groups help a bit [03:19] most of the practical problems we see are timezones and just plain not being in the same room [03:22] cjwatson: so in you're opinion would you say text based communication is a major hindrance to OSS communication and development? [03:22] firefly2442: better shot with text than spoken [03:22] firefly2442: not a lot of the time [03:22] at least text has google translate [03:23] and many of the people around here are polyglots so finding someone that speaks your language and can pass on a "hey i have this patch that fixes ____" messages shouldnt be hard [03:24] i cant think of any programming languages that aren't english-based though, so i would guess if you can program you know at least a few words of english....like integer [03:24] firefly2442: that's a bit of a leap from what I said :-) [03:24] I agree with maco's comments [03:25] ahh ok thanks, I'm just curious because I'm looking into OSS community management for a paper [03:25] certainly I've seen a number of instances where it was very difficult to communicate with somebody in person, but e-mail allowed them to spend more time composing their words [03:25] so in fact I actively disagree that text-based communication is a hindrance [03:26] Yay, tweaking a gconf key and removing the fast user switch applet got some sane options back. Now, how to replicate that for a11y profiles in casper... :| [03:26] good point, people are probably more apt to spend a long time composing a thoughtful email than just saying whatever pops into their head [03:26] well, some people may simply be unable to follow and participate in a spoken conversation [03:27] that requires better English skills than following and participating in a written conversation [03:27] (I generalise a bit but I think this is mostly pretty uncontroversial) [03:28] speaking of timezones: /me -> bed [03:28] cjwatson: night, thanks for the suggestions [03:30] I should probably head off too, thanks all [03:32] haha oh yeah, text communication in a 2nd or 3rd lang is much easier than in person [04:16] Do we have a standard package to substitute when a Debian package depends on locales-all? [04:19] is this back to the thing where locales are trying to bring half of gnome into kubuntu? [04:20] No, this is a package in universe that's depwait for locales-all because we don't have such a package. [04:29] huh, why do we patch out locales-all? [04:30] OTOH, why does any package build-depend on it, I think there's a recipe for generating locales for local use as part of a build? [04:32] phew, finally got a Jaunty VM up-and-running again [04:33] LaserJock: can you IM me? like, the not-IRC way. i'm trying to see what pidgin's doing [04:35] * LaserJock status update-manager [04:35] *stabs [04:35] * maco joins in and stabs pidgin [04:42] * highvoltage <3 pidgin [04:44] slangasek: rmadison locales-all produces a blank stare in response and bgoffice-computer-terms is the package that build-dep on it. [04:45] Since i386 is caught up I was looking for arch all fixes I could throw at the buildd's. [04:49] does the liveCD every use the HD, say a HD that already has a 1GB swap partition?? [04:49] ... or must everything be done in RAM? I'm guessing this one [04:52] \o/ - Just filed a sync request for Spamassassin. It will be the first time since Dapper we've been in sync. [04:54] congrats [04:58] Thanks. [05:02] gnyes itll use existing swap [05:03] grndslm: yes itll use existing swap [05:05] kk... i figured it ought to [05:07] one more question i've asked before, but didn't ask properly... how do various ubuntu repositories work? For example, if I only want to accept security upgrades & not feature upgrades... will the feature upgrades still creep into the security repository?? Downloading the entire package for any upgrade confuses the heck outta me. [05:07] security just has security updates. updates has bug fixes. [05:08] well it's possible that a security update will be newer than whats in updates [05:08] maco: i get that much.... but eventually, they've got to start converging, right? [05:08] but updates isnt features, just bug fixes [05:08] hmm... so the regular repository is where feature upgrades come in? [05:08] Post release, except for a very few exceptions, we don't do feature updates. [05:08] no they dont at all [05:09] We do those in the next development release. [05:09] we dont do version upgrades post-release unless there's a very good reason....like massive amounts of security patches that can only be gotten up changing versions [05:09] interesting [05:09] The one other sort of exception is kernel patches to support newer hardware on LTS releases. [05:10] so if big glaring security problems and bugs require patching such that half the code changes anyway...then an upgrade might happen [05:10] ScottK: when did that start? [05:10] It was done some for 6.06.2 and more for 8.04.x [05:11] It became clear that "LTS, but will only run on two year old hardware" wasn't going to fly. [05:11] haha [05:11] i see [05:12] one last question... inspired by the ubuntu forums... when will ubuntu come out with a rolling release model? =-P [05:12] to heck with separate security and bug-fix repos [05:13] grumpy groundhog? [05:13] that long? [05:13] that's like 23 releases away [05:13] no, thats the name for the mythical "if you use everything as soon as it hits the repos" [05:13] if ubuntu had something akin to debian experimental, that's the name that was chosen. it hasnt actually happened. [05:14] youd have to just go for debian sid [05:14] never thought about that [05:14] AFAIK there are no plans to go Arch-style [05:15] i still like ubuntu's hardware setup better... even more sane defaults than debian. I couldn't get an intel wireless chip to work with debian outta the box... had to wait for ethernet! [05:15] but I'll prolly have to check out debian sid and arch pretty soon [05:15] i couldnt get ethernet to work on windows outta the box...had to wait for sneakernet! :P [05:16] well anyway if intel's firmware was open, we'd be good [05:16] though i do have to wonder how long ago this was [05:16] because i though with the iwl* drivers that binary firmware problem was gone [05:16] i just tried debian 5.0 and no go on my laptop [05:16] oh ok ..that should have iwl, i would think... [05:17] i was pretty sure intel was supposed to have open drivers, which is why I've always bought Intel-based lappies [05:17] yes the drivers are open [05:17] but the firmware isn't? [05:17] its the firmware and i think some sort of firmware-controlling daemon that aren't [05:17] right [05:17] more confusion [05:18] im not sure how much more was opened in the ipw --> iwl transition [05:18] anyway, i'll let you guys get back to work ;) [05:18] lata [05:19] oh yea, about the liveCD issue... if the hard drive only has windows installed, then it can't use the HD, right? [05:21] right [05:21] figures... [05:21] aight, i'm really out this time [05:21] take it e-z [05:30] Did the publisher run last time? It seems like there stuff still "ACCEPTED" that was done well before 3 after. [06:25] is apache2-mpm-prefork the "default" apache package? [06:26] it's the one required for use with libapache2-mod-php5 [06:27] LaserJock: If you need php or a non-threaded apache, go -prefork, otherwise, pick -worker [06:27] I need PHP unfortunately [06:28] well, neither help me out I guess [06:29] Why? [06:29] I'm trying to force either mysql or postresql but apache pulls both in [06:29] so then things get messed up [06:29] hrm? apache does? [06:29] through aprutil [06:30] ah [06:30] aprutil deps on mysql, postresql, and sqlite [06:31] * slangasek tricks upstream into also depending on odbc [06:31] there isn't a way to test what packages are also being installed in postinst is there? [06:31] no [06:31] slangasek: And Oracle, for good meaurse? [06:32] StevenK: sure, why not [06:32] StevenK: just the server though, not the client libs [06:32] I'm sort of close to picking a DB and telling everybody else to go figure it out [06:33] LaserJock: dbconfig-common? [06:33] ah, well, I haven't looked at that yet [06:33] we were using wwwconfig-common [06:33] ah [06:33] but the vast majority of the bugs are due to DB setup issues [06:34] wwwconfig-common suggests: apache | apache-ssl. Vintage. :P [06:34] moodle is often uninstallable in Intrepid [06:34] upgrades seems to be breaking decently often [06:34] is it really appropriate to fix all of this at the same time as fixing the security vulns, though? [06:35] well [06:35] for me personally I'm just trying to get a good, installable version of moodle for Jaunty [06:35] the security stuff is a added benefit [06:36] right now it's uninstallable for most (if not all) people [06:36] so for a Main app I'd think that'd be an issue [06:38] but I'm just a chemist and working with server/webapp stuff is not trivial for me :( [06:40] it sure looks like dbconfig-common would help us out [07:02] LaserJock: but I think refitting it with dbconfig-common is beyond the scope of the merge request that's currently targeted for alpha-5 [07:06] good morning [07:06] hello [07:07] hello dholbach [07:07] hi maco === hunger_t is now known as hunger [08:29] Is there any more firmware/drivers that jockey enables besides Nvidia, Ati, Broadcom, & Atheros... specifically in terms of networking? [08:30] pitti: next time you refresh ubuntu-meta, can you please take care of bug 331256. [08:30] Launchpad bug 331256 in ubuntu-meta "Please remove libpt-1.10.10-plugins-* from ubuntu-desktop dependencies" [Undecided,New] https://launchpad.net/bugs/331256 [08:33] can somebody help me figure out what restricted firmware/drivers jockey enables?? /usr/bin/jockey-gtk doesn't seem to be helping me much [08:35] grndslm: take a look at files in package jockey-common. It should contain one handler for each of the driver category. [08:35] slytherin: sweet... thanks [08:46] soo... it's just b43 (and how is that different from broadcom_wl?), fglrx, nvidia, & sl_modem for software modems?? [08:46] it doesn't even handle madwifi? [08:47] I guess madwifi prolly isn't good enough to be in jockey, tho... which kinda makes sense now [08:48] still... what's the difference between b43 & broadcom_wl ?? [08:48] ahem.. slytherin [08:49] grndslm: I have never heard about broadcom_wl, is that new driver in latest kernel? [08:49] i'm on 8.04.2 [08:50] it's in /usr/share/jockey/handlers [08:52] grndslm, /usr/share/jockey/handlers/broadcom_wl.py exists in intrepid [08:53] and jaunty [08:53] directhex: i wouldn't doubt it... i'm trying to find out how that's different from the b43-fwcutter [08:54] grndslm: I suppose it is just the name that is changed. And probably it has handling for b43legacy and b43. In hardy there was only one driver for all cards if I remember. [08:55] slytherin, /usr/share/jockey/handlers/b43.py also exists on jaunty. why we have two, i dunno [08:55] slytherin: there's b43 & b43legacy in my /lib/firmware/ [08:55] on hardy [08:56] wait, I remember now. broadcom_wl is new driver and it is different from b43. [08:57] grndslm: the best person to ask is pitti. But he seems to be away at the moment. [08:58] i guess i'll just drop him an email and see what he says [08:58] fighting crime from his orbiting skyship, no doubt === mvo__ is now known as mvo [09:07] are any of the notifyosd guys around? [09:08] don't ask to ask just ask? [09:09] who do you call notifyosd guys? you don't accept reply from non notifyosd people who know about what you would ask there? [09:10] seb128: I suppose anyone can answer but I thought they would have a better working knowledge of what was supported and not [09:10] hard to say since you don't ask your question [09:10] we could avoid such pointless discussions if you just ask what you want to know [09:11] sorry for the rant there ;-) [09:11] I'm having an issue with pidgin not triggering notifyosd but I'm not sure if irc is supported by it yet [09:11] no it doesn't work on IRC [09:11] seb128: and in future I'll just ask, thanks dude [09:12] you're welcome [09:12] see it's easier this way ;-) [09:13] also seb128 any idea why it seems to put an envelope in the system tray I thought that was for evo when you got mail [09:13] that's the message indicator [09:13] it's there if you have evolution or pidgin running [09:13] I don't have evo open only pidgin [09:13] and the icon should change when you get a message [09:14] well in this case you should have pidgin listed when clicking on it no? [09:14] * directhex is still sad @ the idea of not having a "next track" button on banshee popups [09:14] yes that's correct :) [09:15] thanks again [09:16] hey by the way, is it a function of the new notification system that pidgin's buddy list is now invisible for me (not in window list or alt+tab) since i manually turned off system-tray-icon or is this a bug? [09:16] Hrm. Where did /etc/adjtime go? I see base-files creates it in its postinst, but none of my jaunty systems have it, and a powerpc alternate install fails because powerpc-utils can't find it. [09:16] maco: is it open on screen? [09:17] seb128: by invisible, i'm not joking. it must exist somewhere...it flickers into visibility for a moment when i quit it. its certainly running since there are notifications and i can receive messages...but the window itself seems to be hidden somewhere that i can't un-hide it [09:18] the notification/tray icon had a "hide buddy list" feature [09:18] maco: did you try to select pidgin the message indicator menu to open it? [09:18] im in kde [09:18] +in [09:18] maco: well you did shoot yourself in the foot apparently then [09:18] hahaha [09:19] maco: ie you turned the notification icon but don't use the message indicator and don't run the right desktop environment [09:20] i get the notification icon for notifying of new messages, however i was under the impression that setting pidgin to not live in the tray meant that it would not attempt to hide, ever [09:20] wrong impression [09:20] >< crap [09:20] seb128: any idea how the envelope should change other than listing the chat and time? should the flap open or something? [09:21] davmor2: not sure, the icon should be different but that might be buggy right now [09:22] seb128: just FYI ... I updated gst-plugins-bad-multiverse to bring in sync with -bad, will update ugly-multiverse sometime this week. [09:22] slytherin: cool [09:23] seb128: I'd of thought that changing it to an open envelope would of made sense for unanswered messages in case your away at the time. [09:24] seb128: I also added some arguments to configure in debian/rules to disable some plugins that get by default but are not included in the package. Reduces build time a lot. [09:24] oh good too [09:32] TheMuso: see util-linux changelog [09:33] Keybuk: ^- powerpc-utils is a difficult one - without having looked I suspect it's coping with known problems on powerpc systems where the firmware time gets reset to the Mac epoch on battery failure, or some such. Do we really need to delete /etc/adjtime, rather than just deleting the code that used it? [09:50] from where is the Default Nautilus-Desktop Background Color loaded if its not set in gconf? [09:53] Silicium: what do you call nautilus-desktop background color? [09:53] it's an image by default [09:54] nautilus/gdm ans so on [09:54] the color i see between GDM and finaly loaded gnome [09:54] this is the "background color" set in the settings [09:54] but i cant find it in the gconf of liveCD [09:55] may is hardcoded? [09:56] Keybuk: so if jaunty is failing to assemble a root LVM after last night's udev/lvm updates, what would you want/need to know? :) [09:56] Silicium: it's a gdm.conf setting [09:56] ok thanks [09:56] that will be changed soon [10:03] asac, why does my FF open a ton of windows after the last upgrade ? (add-ons, the print dialog) [10:05] * ogra could understand add-ons to make him chek if everything is still ok, but why the print dialog ? [10:08] ogra, you don't print a copy of the "your firefox has been upgraded" screen for your wall? [10:08] ogra: dont know about the print dialog [10:08] ogra: the addons dialog opens when there have been addons upgraded/disabled [10:08] directhex, i do ! but the wall is full :P [10:09] asac, yeah, thats what i suspected, but the print dialog is a bit strange [10:09] ogra, phoenix 0.6 has been upgraded! [10:12] ogra: maybe print dialog opened because the "random menu item activates" bug? [10:12] did you click somewhere before it popped up? [10:13] no, i rebooted and started FF, evo and xchat [10:13] ogra: heh. so maybe it wasnt firefox print ;)? [10:13] from the panel where i have shortcut icons [10:13] well, FF came up with it, evo usually takes a while [10:14] and i dont think chat has any print dialog [10:14] * ogra checks [10:14] ogra: ok. let me know if you can reproduce the print thing [10:14] * ogra logs out again [10:16] hmm, add-ons again ... no print dialog [10:16] whoops ! [10:16] mvo: I see no notification icon for update-notifier anymore in jaunty; is this part of the new notification system changes? [10:16] i was wrong, FF blinked in the tasklist after xchat came up, clicking the task entry shows the print dialog [10:16] * ogra tries a third time [10:17] mdz: yes, this is what the design team asked for. no icon, just update-manager starting up automatcially for updates every 7 days (and immediately for security updates) [10:18] mvo: :-( where should I send my feedback? [10:18] mdz: there is a long debate on ubuntu-devel about it [10:18] mdz: there is also a gconf key to get the old behaviour back [10:18] asac, hmm, apparently reliably reproducable ... [10:18] i get the print dialog on every login in FF now [10:19] mdz: and yes, I like the icon too [10:19] andthis time i only started FF [10:19] i also seem to get the add-on dialog every time now [10:19] mdz: seb128 had the idea that we keep showing the icon but auto launch after 7 days if it was not clicked. I like that proposal [10:19] ogra: do you stop your FF sometimes? or always just "log-out"? [10:19] asac, the latter [10:20] i want to keep my session [10:20] without saving it [10:20] ogra: please check your localstore.rdf [10:20] mvo: the restart dialog popping up all the time is also excessive [10:20] ogra: does it a) have proper permissions, b) a recent modification time, c) contain any hints about the print dialog? [10:20] ogra@osiris:~$ ls -l /home/ogra/.mozilla/firefox/x1utrcmr.default/localstore.rdf [10:20] -rw-r--r-- 1 ogra ogra 10296 2009-02-24 11:01 /home/ogra/.mozilla/firefox/x1utrcmr.default/localstore.rdf [10:21] ogra@osiris:~$ grep print /home/ogra/.mozilla/firefox/x1utrcmr.default/localstore.rdf [10:21] ogra@osiris:~$ [10:21] doesnt look like [10:21] ogra: -i ? [10:21] hmm [10:21] nope [10:21] -i doesnt change it [10:21] ogra: what extensinos installed? [10:22] grab n drag and ubufox [10:23] mdz: "all the time"? you probably had a broken upgrade which it's trying to reconfigure and which is breaking again or similar? [10:23] mdz: right, all actions that used to show notifications are now launched directly (including reboot, interactive upgrade hooks etc). some concerns on the ML about this include that stuff that opens out of the blue will remind windows user of spyware/viruses (where this apparently happens too) [10:23] seb128: no, I just haven't rebooted [10:23] ogra: disable g&d? [10:23] and so each time I upgrade, the dialog pops up [10:23] mvo: btw any clue about bug #333557? [10:23] Launchpad bug 333557 in python2.6 "pygobject/python2.6/dpkg upgrade failure" [High,Confirmed] https://launchpad.net/bugs/333557 [10:24] mvo: is there a bug open about this? [10:24] mdz: about the dialog comming up all the time? or the new behaviour in general? [10:24] mvo: either, or both [10:24] should they be considered separate issues? [10:25] seb128: I have a look, but it looks like a simple file conflict? [10:25] asac, no change [10:26] mdz: there is a bug about the new behaviour, give me a sec I look for the number. update-notifier gets bugreports about it all the time basicly, because people think its broken (no icon anymore) [10:26] mvo: I don't think so, look at the path that's not a full one and those are not conflicting [10:26] * ogra wonders if there is any evil javascript on a website that triggers it, nut i didnt have it on yesterdays upgrade [10:26] *but [10:26] ogra: err. you have zillions of tabs? [10:26] and the list of pages didnt change much [10:26] seb128: I suspect it has something to do with the new shared location to install stuff, but i will have a closer look in a bit [10:26] yes, a bunch, but since yesterday only some LP tabs were added [10:26] and i rebooted yesterday as well [10:27] mvo: thanks [10:27] $ dpkg -c /var/cache/apt/archives/python-gobject-dbg_2.16.1-0ubuntu1_i386.deb | grep gobject.so [10:27] -rw-r--r-- root/root 243986 2009-02-23 11:44 ./usr/lib/debug/usr/lib/python2.5/site-packages/gtk-2.0/gobject/_gobject.so [10:27] -rw-r--r-- root/root 246472 2009-02-23 11:44 ./usr/lib/debug/usr/lib/python2.6/dist-packages/gtk-2.0/gobject/_gobject.so [10:27] -rw-r--r-- root/root 333165 2009-02-23 11:44 ./usr/lib/python-support/python-gobject-dbg/python2.6/gtk-2.0/gobject/_gobject.so [10:27] $ dpkg -c /var/cache/apt/archives/python-gobject_2.16.1-0ubuntu1_i386.deb | grep gobject.so [10:27] -rw-r--r-- root/root 122172 2009-02-23 11:44 ./usr/lib/python-support/python-gobject/python2.6/gtk-2.0/gobject/_gobject.so [10:27] -rw-r--r-- root/root 121756 2009-02-23 11:44 ./usr/lib/python-support/python-gobject/python2.5/gtk-2.0/gobject/_gobject.so [10:27] mvo: ^ [10:27] see no conflict [10:29] doko, is our java built with freetype support ? === mvo__ is now known as mvo [10:31] mdz: about the auto launching, bug #331054 and #332945 are probably interessting (both already closed as invalid though) [10:31] Launchpad bug 331054 in update-notifier "Do not launch in background" [Undecided,Invalid] https://launchpad.net/bugs/331054 [10:31] Launchpad bug 332945 in update-notifier "[Jaunty] Removal of Update Notifier is WRONG" [High,Invalid] https://launchpad.net/bugs/332945 [10:31] mdz: there is also the mailing list discussion [10:31] mvo: thank you [10:31] mvo: I'm bringing it to the attention of the design team [10:32] mdz: I think it has there attention already :) [10:32] mdz: i.e. they replied in it [10:37] mvo, do-release-upgrade has no manpage? [10:37] liw: no :( [10:37] liw: I should attend to your manpage write course [10:38] mvo, that needs fixing one of these days... I'll gladly help :) in the mean while, is there a way for me to instruct it to not check for free disk space? my chroot doesn't have /proc mounted [10:39] liw: unfortunately not. could you just mount /proc as a workaround? (mount or bind mount)? [10:39] liw: I plan to add some switches soonish [10:39] liw: do make it possible to have more control, but currently its a bit set-in-its-ways when it comes to that [10:40] mvo, I can mount /proc, no problem, it's just slightly less convenient to have to do that (when creating, testing, and destroying chroots) [10:42] liw: ok, I will make it a option for you later today, ok? [10:42] mvo, nah, don't worry, I've adapted already, no need to hurry this [10:43] apropos the manpage thing: I gave a presentation in Finnish on it, so perhaps I'm ready to give one at the next UDW or something (uds?) [10:43] although I'm sure Colin would hate my style of manpage writing :) [10:43] oops [10:44] the things you do when you're used to using kvm instead of chroot... I let do-release-upgrade reboot, and it rebooted the whole computer, of course, not just the chroot [10:46] * liw waits for the desktop to boot (root is on usb, so it's very slow to boot) and ponders aliasing chroot to kvm [10:49] ogra: what do you mean? java has it's own selection scheme for fonts [10:50] doko, http://sourceforge.net/mailarchive/forum.php?thread_name=892226.3959.qm%40web62005.mail.re1.yahoo.com&forum_name=ltsp-discuss [10:51] doko, see the last mail, we have a longstanding issue with ltsp where java gets unusably slow on remote X connections, appartenly only installing a fontserver solves that, i was wondering if thats solvable in a better way [11:12] everybody welcome rcmorano - he's from Guadalinex and doing great work on Ubuntu too! [11:12] :]! [11:13] * ogra waves to rcmorano [11:13] morning [11:17] Keybuk: tracked it down... for some reason the initramfs had something wrong with its cryptroot hook. re-generated it after doing a manual cryptsetup+boot and it works again. bit weird that that happened though [11:18] cjwatson: thanks [11:19] themuso [11:19] pulse 0.9.15 is a resource hog. [11:19] LLStarks: talk to upstream about it [11:19] LLStarks: tried turning off glitch free? [11:20] eh? [11:20] i'm just using whatever you packed. [11:23] LLStarks: Yes, but I only made that package available for testing, expecting people would go to upstream and report problems. The only thing thats different is a few settings changes to better suit Ubuntu. [11:33] dholbach, the next UDW is sometime during the summer, yes? [11:33] liw: yes [11:35] dholbach, six months from January 19 would be July 19.. do you have exact dates already? [11:35] liw: no [11:35] dholbach, I'm asking so that I can plan my summer so that if possible, I can give the manpage writing tutorial :) [11:35] sorry, no dates yet [11:36] dholbach, no worries, I realize it is very early for that [11:36] :-) [11:38] UDS is may, isn't it? [11:38] directhex, yes [11:39] oh balls..... wife's birthday......... even if i can get sponsorship, i don't know if it's wise to play that particular card [12:00] I've implemented the changes described in https://lists.ubuntu.com/archives/ubuntu-devel/2009-February/027427.html. Please let me know if there are any problems [12:13] seems sound to me, cjwatson [12:22] is new gdm theme is oficial? :) [12:22] grndslm: hey [12:23] grndslm: madwifi is handled, but not with a special handler [12:23] grndslm: I have a bug about that [12:23] grndslm: the b43 handler does use b43-fwcutter, it doesn't reinvent the wheel [12:23] grndslm: so what's your question? [12:25] slytherin: thanks for pointing out === JanC_ is now known as JanC [12:32] is gdm2 going to be in 904? === thunderstruck is now known as gnomefreak [12:33] alex-weej: no [12:34] alex-weej: it's available in the desktop team ppa as gdm-new if you want, might to universe but we have no interest to switch by default, it brings nothing to use and add quite some limitations [12:34] oh ok [12:35] i used it on fedora 10 the other day and thought it actually looked more featured, but i didn't really dig into it [12:36] is there a channel for questions who can answered in #ubuntu? [12:36] so then i dont need to ask it here... [12:36] s/can/can't [12:36] if you mean "which can't be answered", then you could try #ubuntu+1, but if they can't help then that is still not a valid reason to ask here [12:36] Silicium: you could always try http://answers.launchpad.net/ -- that way your questions won't be lost after about 5 seconds when they scroll off the buffer ;) [12:37] if IRC can't help, try mailing lists, web forums, answers.launchpad.net, or a bug report if that's appropriate [12:37] cjwatson: thats to slow [12:37] :/ [12:40] Silicium: I'm afraid that still doesn't justify asking here [12:40] :D [12:41] so, i actually dont want everytime use this chan [12:41] seb128: i take it it breaks FUSA and guest session then [12:42] alex-weej: that yes [12:42] alex-weej: it also has no gdmsetup = no way for user to enable autologin or timedlogin [12:42] alex-weej: it has no settings migration, no theme support [12:42] argh damn shiat [12:43] i see. is it actually even considered ready yet? [12:43] (by upstream) === thunderstruck is now known as gnomefreak [12:44] alex-weej: GNOME ships it since 2.24 [12:45] alex-weej: it doesn't add useful features over the old version out of some bling (fading, etc) but some things are not there yet, we don't think changing would benefit our users [12:46] alex-weej: and we got too many complain about GNOME breaking things (gnome-session-saving, audio, etc) recently, we want a "stable" cycle and not another flameware about new rewrittes dropping useful feature for no user win [12:46] sure, you don't need to defend yourself, i understand :) [12:46] (but saying what you just said... notify-osd seems to have ruffled those exact same feathers) [12:47] alex-weej: I'm just explain the choice not defending myself there ;-) [12:53] pitti, seems there are a few niggles with apport suspend/resume still. i've pushed up a couple more fixes, merge proposal here: https://code.launchpad.net/~apw/apport/suspend-resume/+merge/3899 [13:11] mvo, seb128: any clues about the python-gobject-dbg/dpkg conflict? [13:11] doko: no [13:11] I'm pretty sure that's not a python-gobject issue though [13:12] doko: I just ran it through my auto-upgrade tester and did not get it there [13:12] nothing special changed there and my local build didn't have the issue [13:12] mvo: do you have an uptodate jaunty, does apt-get install python-gobject-dbg works? [13:13] do you have python2.6 installed as well? [13:14] doko: you have to since the current pygobject depends on it no? [13:14] let me check via chroot [13:14] cjwatson: it's only deleted if it's not being used [13:14] since otherwise it gets written with increasingly random values on shutdown [13:15] cjwatson: I didn't know base-files created it, it shouldn't [13:16] apw: thanks, merged; running test suite now [13:16] hi people, is there any plans to get rid of evolution-indicator dialog thingy in Jaunty? :) it drives me m(a)*d [13:17] pitti, thanks ... sorry for so many iterations, this thing is sooo hard to test properly without crashing kit every time [13:17] apw: no problem at all :) [13:17] apw: erm, you don't have a "fake" susres.crash for testing the UI? [13:18] pitti, ? [13:18] apw: I mean you actaully have a machine where suspend fails, and do that to test the GUI? [13:18] seb128: no, doesn't. but it depends directly on python2.5, which it should not [13:19] pitti, na, i can fake everything to test the UI component [13:19] Keybuk: doesn't it only get written on shutdown if you turn on that code? or is that unconditional? [13:19] but as i found with the last iteration if you don't actually crash it [13:19] cjwatson: it was unconditional in the init script [13:19] so I made it conditional on the existance of /etc/adjtime [13:19] you don't really run it quite the way that a crash does ... [13:19] Keybuk: how do you propose to fix the powerpc problem? adjtime was actually a useful "this is a reasonable minimum time that it could possibly be" [13:19] cjwatson: what _is_ the powerpc problem? [13:19] doko, should have openjdk-6 finished building on ARM? It's been going for a week, and I think it might be stuck in an infinite loop (the Debian build only took 3 days ...) [13:19] apw: right, you should keep a real .crash file around and just cp it to /var/crash for UI testing [13:19] Keybuk: I described it earlier [13:20] cjwatson: i don't have an e-mail? [13:20] (or a bug report) [13:20] oh, hmm, maybe I misdescribed it though [13:20] 09:33 Keybuk: ^- powerpc-utils is a difficult one - without having looked I suspect it's coping with known problems on powerpc systems where the firmware time gets reset to the Mac epoch on battery [13:20] apw: I'll do another fix, and then upload [13:20] failure, or some such. Do we really need to delete /etc/adjtime, rather than just deleting the code that used it? [13:20] Keybuk: you were replying to me so you clearly saw it :-P [13:21] I just see the bits you say [13:21] pitti, cool [13:21] I don't see how adjtime solves that [13:21] adjtime is for dealing with drift between the two clocks [13:21] not one clock always saying 1 Jan 1970 [13:21] base-files updates adjtime every upload to a reasonable minimum time [13:21] NCommander: which Debian build? [13:21] I think Santiago just updates it to the current time or thereabouts [13:21] we actually don't have anywhere else to say "hey guys, it's at least 2009" ;-) [13:22] 09:16 Hrm. Where did /etc/adjtime go? I see base-files creates it in its postinst, but none of my jaunty systems have it, and a powerpc alternate install fails because powerpc-utils can't find it. [13:22] cjwatson: that doesn't fix the problem though [13:22] doko, openjdk-6 on Debian on armel took three days. The last upload to Ubuntu has been going for almost a week, and has been spitting out the message "Compiler still running" for most of that time :-) [13:22] and would cause others, since suddenly hwclock would be using adjtime again [13:22] if it is based on the same package, then it looks like a problem. but imo these are the arm specific optimizations [13:22] for everything [13:22] it doesn't *fix* it, but sometimes a vague approximation of recent time is better than 1904 [13:22] NCommander: which version was the Debian build? [13:22] (which is the Mac epoch, and happens to be negative Unix time, which breaks some things in fun ways) [13:23] doko: 6b11-9.1 [13:23] Keybuk: calm down, I'm not saying "argh we must reintroduce adjtime" :-) [13:23] one obvious solution is ntpdate ;) [13:23] Keybuk: I think I may have misdescribed the powerpc-utils problem, though; it looks to me as if it ships an hwclock port [13:23] Keybuk: one of the breakages caused here is that the desktop refuses to start :-P [13:24] NCommander: that's some months old ... [13:24] anyway, I suspect what we actually need to do here is port powerpc-utils' clock to the new world order [13:24] or figure out whether it can be thrown away in favour of hwclock [13:24] right [13:24] if we need minimum time support, that can be patched into hwclock itself [13:25] maybe it could use the timestamp of its executable as a minimum time or something [13:25] doko, seb128: I have the error now, I look at it [13:25] mvo: thanks [13:25] cjwatson: that's actually not bad idea ;) [13:25] its pretty clearly a python-support issue it seems :/ [13:26] anyway, I think minimum time was a red herring in response to Luke's actual bug - I hadn't looked at the code yet when I brought that up [13:26] so sorry about that [13:26] mvo: I would guess so too [13:26] doko, I was asking if these kind of build times are normal. openjdk-6 uploads on ARM have taken two days [13:26] doko, so I'm starting to wonder if something has gone horribly horribly wrong [13:27] *sigh* [13:27] yeah for two implementatons of the same [13:27] mvo, hmm, do-system-upgrade -d removes system-cleaner(-gtk), but doesn't install computer-janitor -- when I tested this earlier with apt-get dist-upgrade, it worked; should I introduce a transition package? [13:27] powerpc-utils.postinst copies /etc/adjtime into /var/lib/hwclock/, apparently due to Debian #280605 [13:27] Debian bug 280605 in powerpc-utils "/sbin/clock: Newer FHS require adjtime in /var/" [Normal,Closed] http://bugs.debian.org/280605 [13:27] and that's the thing that's actually failing [13:28] heh [13:28] btw. can base-files be made to not create /etc/adjtime on non-ppc machines? [13:30] seb128: python2.5 dependency is hardcoded in control.in. please remove it for the next upload [13:30] same for the -dev package [13:31] doko: ok [13:31] Keybuk: in principle sure but I suspect there is really no reason not to fix powerpc [13:32] cjwatson: what I mean is, if base-files is creating it, then hwclock will use it unconditionally [13:32] even on x86 [13:32] that's the reason it gets removed in the hwclock postins [13:32] to stop it being used [13:32] right, I'm just questioning the "on non-ppc machines" bit [13:32] ah ok ;) [13:35] cjwatson: You have a local mirror, right? Does it also mirror all the installer bits? I can't seem to get debmirror to grab enough stuff to make jigdo happy. [13:36] yes, it does [13:36] ttp://paste.ubuntu.com/122373/ [13:36] http://paste.ubuntu.com/122373/ [13:37] cjwatson: Aha! [13:41] seb128, doko: http://paste.ubuntu.com/122375/ - so now its just "lets fix pysupport (or switch to python-central ;)" [13:43] mvo: now quick, seb128 just left, lets switch to python-central ;) [13:43] cjwatson: That doesn't grab updated installers in -updates, does it? [13:43] soren: nope [13:44] cjwatson: Alright. Thanks! === chand_ is now known as chand === __iron is now known as iron [13:54] * Keybuk has coffee now [13:55] * cody-somerville has cranberry juice. [13:55] * directhex has savaged his fingertip with a pair of pliars [13:56] * Laney cannot figure out why n != S n' [13:57] * Laney has to go mark first year undergraduate presentations now. wish me luck [14:03] slangasek, can you poke the mythbuntu daily builds so we can get to checking for alpha5 rc? it looks like a hash sum mismatch again from apt-get update again (http://people.ubuntu.com/~ubuntu-archive/cd-build-logs/mythbuntu/jaunty/daily-live-20090224.log) [14:05] slangasek, er well hold off i think. looks like there is something wrong with clock-setup in the current ubiquity anyhow, so we'll need to wait for one more at least [14:19] Keybuk: how come DM_TYPE can be empty? broke multipath here :) [14:20] tjaalton: no idea? [14:20] do you mean the table type? [14:20] you can create devmapper devices with no table [14:20] Keybuk: it used to work until I reinstalled the box today. kpartx rule doesn't finish [14:21] device-mapper table, I guess [14:21] I don't think anything's changed there recently? [14:21] not there, but in udev? [14:23] Keybuk: how did I get the debug info again, from busybox [14:25] I mean the options for udevd [14:30] tjaalton: --debug [14:41] Keybuk: ok, DM_TYPE was exported by kpartx_id, and not having it in the image doesn't help :) [14:41] s/was/would be/ === Koon is now known as ttx === ttx is now known as ttx_ === ttx_ is now known as ttx [15:07] Keybuk: that didn't help much, the rule is still not finished. I suspect this has something to do with the inotify changes in udev/devmapper [15:07] is there a way to tell if a package is a standard seed. i want to know if checkbox is a default install on desktop and server [15:08] tjaalton: rule is still not finished? [15:08] Keybuk: kpartx.rules [15:08] oh, that's multipath-tools crack? [15:08] has that _ever_ worked? [15:08] yes [15:08] yes [15:08] :) [15:08] why isn't it "finished" ? [15:08] apw: apt-cache show checkbox, look for the Task field [15:08] so those are the tasksel tasks? [15:09] Keybuk: you tell me :) [15:09] I'd like it to reach the 'kpartx -a.. ' phase [15:10] but it doesn't [15:10] I don't know, i've never read that rule [15:10] if it doesn't _reach_ something, then something before it is taking over three minutes [15:10] it worked fine until the reinstall [15:11] apw: yes [15:11] apw: those are constructed by dependency-expanding seeds [15:12] cjwatson, rocking thanks ... yeah seen those in the germinating phase for CD creation, terifying [15:13] is mvo responsible for the gdebi source package? or should I subscribe ubuntu main sponsors? there's a patch for bug 190907 [15:14] Launchpad bug 190907 in gdebi "[kde] Applications cannot read Greek folder names" [Undecided,In progress] https://launchpad.net/bugs/190907 [15:15] savvas: if it's for the KDE side you can poke me (subscribing ubuntu main sponsors a good idea too) [15:15] savvas: thanks, you are quick :) I know about the patch, I think its good, but I need to look a bit closer. but if Riddell could have a look that would rock of course [15:16] Riddell: if you think its fine, I apply it, it looks good to me, but I know much less about pykde than you :) I tried to contact martin.boehm too [15:16] Riddell: ok, noted :) [15:18] thank you both for reviewing it :) [15:37] \o\ \o/ /o/ dholbach! \o\ \o/ \o/ [15:37] highvoltage: hm? :-) [15:37] sorry, just couldn't keep it in [15:37] hehe [15:37] * dholbach hugs highvoltage [15:38] there's one compelling reason to go to UDS [15:38] booze? [15:39] dholbach-hugging! [15:39] ice cream eating comp [15:39] :-D [15:39] mmm ice cream [15:42] pitti, you still around? [15:42] grndslm: hi [15:42] pitti: heya... whole reason i'm asking you about jockey drivers is because i'm trying to use remastersys to create a distro with [at a bare minimum] all the restricted drivers [15:43] soo... i know i need to install madwifi, b43-fwcutter, and is there anything else? main question in the email is what's the diff between b43-fwcutter and broadcom-wl [15:45] pitti: any hints? [15:47] grndslm: I got your mail; do you prefer IRC or email answer? [15:47] honestly doesn't matter... irc since we're already here ;) [15:47] grndslm: b43 and broadcom-wl are two totally independent drivers [15:48] grndslm: wl is shipped in l-r-m (including firmware), b43 is in the main kernel (free module), but needs non-redistributable firmware (through b43-fwcutter) [15:48] grndslm: so you should ship either [15:48] grndslm: jockey also provides installation of nvidia and fglrx drivers [15:48] pitti: i'm mostly worried about wireless for now, so then jockey can be used to setup graphics later [15:49] pitti: what is l-r-m? [15:49] grndslm: linux-restricted-modules [15:49] linux-restricted-modules [15:49] ahh... so if i have l-r-m on hardy, i've already got the wl driver? [15:50] grndslm: I think it was added to linux-backports-modules in hardy; please check the package contents and changelog [15:50] could also have been l-r-m, though [15:54] pitti: how do i change package contents & changelog? =D and is it possible to install both graphics drivers and have the LiveCD just automatically choose one *after* installation? [15:54] *check pkg contents [15:54] grndslm: dpkg -S /path/to/installed/file gives you the package which ships a filel [15:54] grndslm: no, you can't *install* the graphics drivers, you can at most ship them [15:55] soo.. just have the debs already downloaded for both of them? [15:55] first, the 4 nvidia drivers conflict to each other file-wise, and second they do a lot of fiddling with the OpenGL libraries, etc. [15:55] sounds about right [15:55] same with fglrx [15:56] you really only want to install this on systems which actually use fglrx [15:56] 4 nvidia drivers? you mean different versions of the same nvidia driver? [15:57] any nvidia version should work on all gfx cards, eh? [15:58] grndslm: right; unfortunately they all support a different set of models [15:58] grndslm: unfortunately not; if that were so, we wouldn't have four in the first place [15:59] wow... never realized that [15:59] pitti: is this the same wl driver we're talkin' about? http://pastebin.ca/1345923 [16:00] grndslm: that one, yes [16:00] it's in linux-restricted-modules... so i guess i'm all good to go in the networking department then? [16:01] pitti: no more wireless drivers i need to add? [16:02] grndslm: we don't have other packages related to that, no [16:03] grndslm: you might want to include ndiswrapper [16:03] grndslm: jockey doesn't support it, but it's not terribly hard to set it up manually, or with ndisgtk [16:03] pitti: it's useless without firmware tho isn't it? [16:03] grndslm: s/firmware/a windows driver/ [16:04] but it provides the infrastructure to use the bits from windows driver CDs [16:04] * pitti never used it so far [16:04] i'll never understand the difference between the two [16:04] 3v1l [16:04] cody-somerville: how does one check/set the keyboard layout in XFCE? [16:04] grndslm: firmware runs on the device, a driver runs on your computer [16:04] i guess i should just install it incase somebody actually has their driver cd [16:04] grndslm: well, "driver" usually encompasses firmware, too [16:04] Keybuk, in Jaunty? [16:04] cody-somerville: in general [16:05] In Jaunty, Applications > Settings > Keyboard. Click the layout tab. [16:05] pitti: i get that part, but how does downloading broadcom firmware & stripping it lead to a linux compatible solution? [16:05] cody-somerville: where is "Applications" ? [16:06] I have something that looks like a spanner on the task bar [16:06] Keybuk, Same place it is in Ubuntu - upper left corner. [16:06] that has Settings [16:06] and has Keyboard Settings [16:06] but that has no "Layout" [16:06] this is not Xubuntu, btw [16:07] What version of Xfce is it? [16:07] cody-somerville: 4 [16:07] 4.4.3 [16:07] Keybuk, You might have to add the keyboard layout switcher applet in 4.4 [16:07] pitti: thanks for the help, man!! you're too good. [16:08] grndslm: my pleasure, good luck with your project [16:08] but i gotta go for now, so ttyl [16:08] thanks! [16:09] cody-somerville: I can't see such a thing in the list [16:10] Keybuk, xfce4-xkb-plugin is the ubuntu package [16:10] cody-somerville: this isn't Ubuntu ;) [16:11] Well, I don't think 4.4.x supports that really. You could use xfce4-xkb-plugin but you'd have to define the list of layouts in xorg.conf. [16:12] there's no xorg.conf here ;) [16:12] Keybuk, The latest version of xkb-plugin can use libxklavier [16:12] and there is always xfkc [16:13] I don't really want to change this much [16:14] Keybuk, Then your only other option is to replace the actual keyboard :P [16:14] cody-somerville: lol [16:14] actually, I want to test whether moblin even lets you set a keyboard [16:14] because I don't think it does [16:17] seb128: chpe says r3325 of gnome-terminal fixes the memory corruption [16:17] gnome bug 572549 [16:17] Gnome bug 572549 in general "Memory corruption in gnome-terminal" [Critical,Resolved: fixed] http://bugzilla.gnome.org/show_bug.cgi?id=572549 [16:17] dholbach: you have upload right feel free to backport the change ;-) [16:18] I'm too busy to work on that today [16:18] and we have a meeting in 11 minutes now [16:18] dholbach: but thank you for keeping track of the issue ;-) [16:18] seb128: no worries, I'll do it [16:18] dholbach: if nobody backport it next tarball are due in a week [16:19] right-o - I just want to try it myself and if it helps to save the work of others, it might be worth it :) [16:19] thank you [16:20] seb128: you're lucky - gnome-terminal does not use quilt, so I'll add the patch ;-) [16:20] dholbach: he he, it might be using bzr though since mvo touched it so be carreful ;-) [16:21] seb128: you'Re right, it does [16:23] seb128: which one (sorry, my network hates me) [16:25] mvo: g-t [16:28] yep, bzr! [16:32] Riddell: did you had a chance to look at the patch? [16:32] mvo: not yet === allquixotic_ is now known as allquixotic === dholbach_ is now known as dholbach [16:38] liw: I commited the skip-free-space-check [16:43] slangasek, pitti: what do you think about updating shared-mime-info to 0.60, it adds a bunch of mimetype definitions [16:44] seb128: what do these do? for opening files with programs? [16:44] the debian version also switch to use dpkg triggers, not sure if you consider that as a risky change after feature freeze or not [16:44] seb128: triggers are well understood by now, so they don't scare me [16:45] pitti: that's a mapping between file content and mimetype, ie "java" on the first line is a java source, etc [16:45] pitti: editor /usr/share/mime/packages/freedesktop.org.xml [16:45] ah [16:46] seb128: well, independently of FF we should review the diff for sanity, other than that it sounds okay [16:46] i. e. to not point to programs we don't install/have/don't want to support, etc. [16:47] pitti: http://paste.ubuntu.com/122452/ [16:47] pitti: there is no program mapping, that just define the mimetypes, then the mapping in .desktop for each applications which list the mimetypes they can handle [16:47] pitti: the diff is the revelant part, ie the database, translations etc should not be an issue [16:48] pitti: do you know much about HAL and keymaps? [16:48] seb128: looks fine to me [16:48] Keybuk: enough in order to be able to fix stuff and commit fdi updates upstream [16:48] pitti: basically I'm trying to work out how the X keyboard map is decided/set [16:48] it's not in /etc/X11/xorg.conf as I would expect [16:49] pitti: thanks [16:49] seb128, mvo: do you use ~ubuntu-desktop or ~ubuntu-core-dev? [16:49] but after that, I'm not sure I know [16:49] Keybuk: that's actually the xorg -evdev driver, which translates the kernel key symbols (linux/input.h) to XF86... key names [16:50] dholbach: whatever is in the control [16:50] pitti: how does it know that I want the "gb" keyboard map? [16:50] seb128: very good point :) [16:50] seb128, mvo: filed a merge proposal [16:51] dholbach: thanks [16:51] Keybuk: I don't think X.org uses its own layout configuration any more; it reads it from hal nowadays, IIRC (tjaalton should know better) [16:51] pitti: right, but how/where/etc. [16:51] Keybuk: /usr/share/hal/fdi/policy/10osvendor/10-x11-keymap.fdi [16:51] that's exactly what I'm trying to trace [16:51] Keybuk: (sorry, we are in meeting, I'm lagging) [16:52] ah, and how does that differ to 10-keymap.fdi ? [16:53] Keybuk: there I'm leaving familiar terrain (bryce/tjaalton), but I *think* 10-keymap.fdi adds the hal-setup-keymap callout to input devices (which means to call setkeycodes to assing scan codes to the linux key codes, defined in hal-info) [16:53] Keybuk: and 10-x11-keymap pokes the keyboard layout into X11 === zaafouri is now known as zaafouri` [16:54] ok... [16:54] Keybuk: i. e. it reads it from console-setup and writes it into input.xkb.{model,layout,variant,options} [16:54] so what in all this calls xkbcomp to actually compile the keymap? [16:54] which is where X is reading it from if you select "evdev keyboard" in GNOME [16:54] you can also set another layout in GNOME, though [16:54] Keybuk: I don't know, I'm afraid [16:54] dholbach: if you give me the url I merge now [16:55] tjaalton: don't suppose you know? [16:56] it looks like it's something in the X server that does that to me [16:57] Keybuk: hal picks it up from /etc/default/console-setup [16:57] mvo: lp:~dholbach/gnome-terminal/fix-memory-corruption [16:57] tjaalton: right; itùs the bit qfter thqt Iù, trying to trqck dozn [16:57] whoah, that was an interesting one ;) [16:57] *chuckle* [16:58] something somewhere must call xkbcomp to get a compiled xkb map to load into the X server [16:58] and I'm wondering where that thing is [16:58] and what it does with the output === fader_ is now known as fader === murdok_ is now known as murdok [16:59] hmm, the X server itself apparently [16:59] 5626 25049 5626 5626 tty7 5626 S+ 0 0:00 sh -c "/usr/bin/xkbcomp" -w 1 "-R/usr/share/X11/xkb" -xkm "-" -em1 "The XKEYBOARD keymap compiler (xkbcomp) reports:" -emp "> " -eml "Errors from xkbcomp are not fatal to the X server" "/var/lib/xkb/server-0.xkm" [17:00] but that named file doesn't exist at the end [17:01] Keybuk: ok, i've not dug that far [17:02] tjaalton: any luck with your kpartx problem btw? [17:02] there's a recent patch to use cached files after the initial run, so it should speed things up [17:02] tjaalton: that's the very patch that breaks the world for me ;) [17:03] Keybuk:hah :) [17:03] tjaalton: do you have a reference to the patch so I can check it's the same one I'm looking at === beuno_ is now known as beuno [17:05] Keybuk: it was sent to xorg-devel@ on feb 19th [17:05] pitti: poppler 0.10.3 to 0.10.4 -> http://paste.ubuntu.com/122462 if you have a minute to review [17:05] by dan nicholson [17:06] seb128: just bug fixes, go ahead [17:06] MacSlow: did your git been moved from f.d.o ? [17:06] pitti: thanks [17:06] gaspa, I hope not [17:07] MacSlow: http://cgit.freedesktop.org/users/macslow/cairo-clock/ tell me "no repo found" [17:07] gaspa, I've not touched it since the ssh-incident [17:07] how do you get a backtrace when the program having the problem catches its own error? [17:07] tjaalton: hmm, this is quite different from the one I have [17:07] ahh, this just avoids reworking if you change the keymap from "gb" to "gb" [17:08] http://people.ubuntu.com/~scott/xkbcomp.patch [17:08] Keybuk: about kpartx; I'll continue debugging it tomorrow [17:08] calc: ideally, fix the program to re-throw the signal afterwards [17:08] is the one I'm playing with [17:08] calc: or gdb it and set a breakpoint in the error handler, perhaps? [17:08] pitti: ok, the program in this case is dselect :) [17:08] i'm currently on my e71, so putty can't open links :) [17:08] gdb just gives me a "Program exited with code 02. [17:08] *blows the dust off*, oh, dselect [17:09] Keybuk: ^ [17:09] i'm seeing bug 252001 [17:09] Launchpad bug 252001 in dpkg "dselect: failed to create baselist pad" [Undecided,Confirmed] https://launchpad.net/bugs/252001 [17:09] actually i could probably just grep the source to find out why that is happening, heh [17:11] MacSlow: ah! it works with: http://cgit.freedesktop.org/~macslow [17:11] ah i found what appears to be causing it [17:11] the ddebs repo [17:12] gaspa, ok === fader is now known as fader|lunch === RainCT_ is now known as RainCT === slangasek changed the topic of #ubuntu-devel to: Archive: feature-frozen | main frozen for alpha-5 | Ubuntu 8.10 released! | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs | LP believed fixed - please report any further timeouts to #launchpad [19:14] hi [19:14] hello [19:15] how can I get a list of the packages installed by default on a normal ubuntu system? I'm building popcon2, so it would be great to have it, so we can filter the 'default packages' and let the other packages show :) [19:17] hmm, not sure of a trivial way to get that information. By "normal" ubuntu system, do you mean a desktop system, or any install (including Kubuntu, Ubuntu server, etc)? [19:22] well, any system that can have popcon enabled, but mainly K/X/Ubuntu === fader|lunch is now known as fader [19:37] zul: was there an upstream bug number for 330626? (no upstream bug task linked) [19:37] zul: wondering if it happens to be fixed in the 3.3.1 release [19:37] yeah gimme a sec [19:38] slangasek: https://bugzilla.samba.org/show_bug.cgi?id=6126 [19:38] bugzilla.samba.org bug 6126 in Printing "Segfault when loading printers." [Major,New] [19:38] * slangasek links it [19:39] so is it now not possible to logout/reboot from the menu? [19:39] LaserJock: correct; if you have FUSA, FUSA is the single place to logout/reboot [19:39] instead of having menu redundancy [19:39] LaserJock: ctrl+alt+del stil works, too [19:39] and ctrl-alt-prtscr-k [19:40] wait, no, that's something different [19:40] slangasek: that one doesn't work for me, baad things happen [19:40] well, closing the door vs. blasting it out :) [19:40] so what if I don't have FUSA? [19:40] LaserJock: you'll get the menu entreis [19:40] entries, too [19:40] oh, nifty [19:41] even dynamically [19:41] I don't use the FUSA applet, but I usually leave it around just 'cause it's default [19:43] that's pretty cool that the menu changes instantly [19:50] apw: I've been following discussion on the kernel-team list about pulseaudio glitch-free vs. kernel configs. I tried your lp276476 PPA kernels (running Jaunty on top of it) and my glitch-free problems are not resolved. Failing that, do you think it's reasonable to resurrect the lowlatency kernel flavor which evidently died in Feisty? === rickspencer3 is now known as rickspencer3-afk [19:52] allquixotic, is not the issue that glitch-free is simply not glitch-free in meaning, it means trust we have no glitches and take longer before trying to fill the buffer [19:52] so to get glitch-free sound you turn off glitch-free mode [19:54] glitch-free means the glitches have been set free, obviously ;) [19:54] apw: Yeah, it's interesting that glitch-free does the opposite :) Are we planning to turn off glitch-free by default then? With compositing enabled by default on my Intel GM965, whenever a big GEM request goes through (such as minimizing a window causing that slide animation), glitch-free glitches :) [19:55] yeah as i understand glitch-free it means we wait until the buffer is nearly empty before refilling instead of the more normal fill any time there is any space, so it reduces cpu usage [19:55] allquixotic: see dtchen's mail to ubuntu-devel on 19 Feb about this [19:55] It just seems like 9.04 should ship with some method of addressing the problem of "I have a driver that makes kernel latency bad" [Intel GEM, Nvidia, others...?] leading immediately to "sound clicks and pops". [19:55] so we can either make the kernel more responsive or turn off glitch-free as you said [19:55] either way resolves my problem [19:55] the latter seems less scarey [19:57] apw: Yeah, as a kernel release manager I would also be somewhat scared of shipping CONFIG_HZ=1000 and a preemptible kernel as the default from an Ubuntu Desktop install. Of course, you could resurrect lowlatency and make it an advanced option in the installer.... but.... too late for 9.04, probably :) [19:57] (I am not, of course, a kernel release manager, but in your shoes, I can see the concern) [20:00] well, there is an -rt kernel in the archive [20:00] allquixotic: dtchen said there's a secondary problem with glitch-free related to a pointer (I'm trying to remember the conversation) and it affects all drivers [20:00] ... whose meta packages need updating :P [20:00] or...no hang on. now im not sure if that was about glitc-free or auto-spawn === RainCT_ is now known as RainCT [20:01] slangasek: I noticed the bug with the meta-package today; I installed `linux-rt` and balked at the version, then aptitude search made me even more confused :) [20:01] he said ideally glitch-free & auto-spawn are both working and shipped, but he wants them on in development so theres at least a way to find the bugs to start fixing them even if they dont ship [20:01] because it exposes bugs in the underlying drivers [20:02] and he'd like to get the drivers fixed [20:02] Lennart has a pretty solid effort going on between him, the ALSA core devs, and the community with his test scripts... he's trying to track down hardware-dependent bugs that cause weirdness in glitch-free. That's a different subject than latency; some hardware can't properly use glitch-free even with excellent latency [20:03] if we wanted to get the benefit of that work, we would have to pull in ALSA from 1.0.19 + git patches that will probably result from the findings about five different ALSA drivers that report bad timing information [20:04] that's still in the works though [20:04] PA 0.9.15 also fixes a bunch of what's wrong in 0.9.14, but of course it introduced another boatload of new bugs [20:05] and 0.9.15 would require doing what you just said with alsa 1.0.19 [20:05] well really, itd probably require just plain going to 1.0.19 overall [20:05] maco: Heh, 0.9.15 has a major bug currently in git, at least with our libtool (not sure if Lennart caught it with Fedora's libtool). It mysteriously can't find the module-alsa-card.so module, which is crucial [20:06] .... i have 0.9.15 installed [20:07] maco: 0.9.15 isn't released yet :) [20:07] well TheMuso's PPA of it [20:08] his PPA probably hasn't pulled in the bug from git :) [20:08] was just about to say "maybe he hasn't built recently" [20:09] I'm trying to track down the root cause of the bug to get it fixed just in case we decide to update PA or ALSA before 9.04 [20:09] ok === asac_ is now known as asac [21:24] Question about bug #216761. Intrepid currently has 3.3.0-1ubuntu7. I got 3.3.0-1ubuntu8 uploaded into Jaunty. Do I need to construct a 3.3.0-1ubuntu7.1 for an Intrepid SRU, or can 3.3.0-1ubuntu8 be used? [21:24] Launchpad bug 216761 in xen-3.3 "errors in xendomains init script" [Undecided,Confirmed] https://launchpad.net/bugs/216761 [21:24] 7.1 === maco_ is now known as maco [21:30] Anybody willing to upload http://launchpadlibrarian.net/23083056/xen-3.3_3.3.0-1ubuntu7.1.debdiff to intrepid-proposed? :) [21:31] Oh crap - don't upload that one, actually [21:35] Ok - let's try this again with a patch going in the right direction. Anyone willing to upload http://launchpadlibrarian.net/23083113/xen-3.3_3.3.0-1ubuntu7.1.debdiff to intrepid-proposed? [21:35] bryce: would you mind if I upload the fix for freedesktop bug 19304, or do you prefer to wait until upstream commits the patch? [21:35] Freedesktop bug 19304 in Driver/intel "[i945 FBC] spontaneous black screen (major pipe-A underrun)" [Normal,Reopened] http://bugzilla.freedesktop.org/show_bug.cgi?id=19304 [21:35] bryce: I had several people test it successfully (also for intrepid, in my ppa) [21:36] pitti: looking [21:37] pitti: sure go ahead. [21:37] pitti: if upstream ends up committing something substantially different, we'll need to make sure to update [21:38] I'm not certain that we will be pulling -intel 2.7 for jaunty. Intel releases are sometimes not so stable [21:40] by "spontaneous black screen" does that mean goes black and stays black, or does it mean a black flicker when it redraws? === rickspencer3-afk is now known as rickspencer3 [21:45] oh ok nevermind...was looking for the launchpad bug :P [21:47] zul: as a data point, I've just found that I can't reproduce bug #330626 on Debian sid with 2:3.3.0-3; I don't think it's a difference in anything being patched in the source, so maybe a difference in build options or compiler settings [21:47] Launchpad bug 330626 in samba "9.04 Jaunty Samba 2:3.3.0-3ubuntu2 fails to start on update " [Unknown,Confirmed] https://launchpad.net/bugs/330626 [21:48] or a difference in the existing environment for each of the servers - will try to narrow this down [21:49] slangasek: well, I've got a merged moodle package that works as well as the current Jaunty one does [21:49] LaserJock: that sounds like an endorsement to me ;) [21:49] slangasek: but considering that the current one doesn't install/remove that's not saying much ;-) [21:49] heh [21:49] apparently it works fine on Debian [21:50] but something's going wrong in the DB creation in Ubuntu [21:50] I don't think I'm going to get all that figured out by Thursday [21:51] so should I bump the milestone or upload the merge now and fix the DB issues after? [21:51] if it's not installable, is it also not upgradeable? [21:52] it might be upgradable if you have an existing working install [21:52] the problem it's having is in the initial DB creation [21:52] if you already have a DB it might work just fine [21:52] but I haven't tested it as I don't have a working DB [21:53] is the intrepid package also broken this way? [21:53] similarly [21:53] oh, intrepid's at the same version [21:53] I'm not sure how far it's broken in Intrepid but I have severall "can't install/upgrade" bugs for Intrepid [21:53] how far back does one have to go to find a version of moodle that's actually installable? :P [21:54] Hardy I guess [21:54] LaserJock: got a package somewhere to look at? I'll be needing to get familiar with moodle soon [21:54] Intrepid's seemed to work if you choose the right options [21:54] but the default DB of postgresql doesn't work [21:54] LaserJock: but intrepid and jaunty are at the same version, so "choose the right options" should also work in jaunty? [21:55] in which case, I would say it's worth having the security updates in now and the maintainer script fixes in later [21:55] assuming that it's just moodle that's the problem [21:56] I think it's a moodle+DB issue so it's possible that Intrepid's moodle+mysql works but Jaunty's moodle+mysql doesn't [21:56] for instance, I haven't had time to go through all the permutations [21:56] ajmitch: Debian has a git repo [21:57] right now the package I have should be functionally as good as current Jaunty's but with security and other bug fixes [21:57] and everything you've merged is there? [21:57] and an internal lib removed [21:57] ajmitch: yep, master branch is what I'm working on [21:58] I just have a heavy load this week so I need to know if what I've got is OK for Alpha5 or not [21:58] the Debian maintainer said he'd help me look into the DB creation issues, but I doubt it'll be resolved before Alpha 5 is out [21:59] but we are merged [22:47] zul: just reproduced the problem with 3.3.0-3 by building with -PIE [22:47] (the Debian package calls ./configure --disable-pie) === murdok is now known as pablog