[00:11] is there a way to break a rules run via a command, eg exit 1 or something? [00:11] if it finds a problem to force the build to fail? [00:13] calc: as in debian/rules ? [00:29] Keybuk: yea [00:29] so i guess something to throw an error for a makefile [00:29] calc: then yes, you can just exit 1 [00:30] ok thanks :) [00:30] i didn't know if there was a more correct way to do it [00:30] heya [00:30] emgent: hi [00:30] calc: heya :) [00:31] i'm adding a test to make sure i stop forgetting to apply the Desktop renaming patches for OOo [00:31] forcing it to kill the build if i do :) [00:32] hard work for OOo :ç [00:35] calc: at the beginning of the build, I assume? :p [00:36] yea [00:39] you wouldn't want to wait a week to find out [00:45] mneptok: 1 hour 15 minutes, come and get your membership! [00:49] cgrega1: 1 hour and 10 minutes until membership meeting! [00:49] awalton__: same for you ^^ 1 hour 10 minutes [00:53] nixternal: Yep..... [00:53] nixternal, thanks. [00:54] no prob [00:55] Hey all [00:57] Is there know problem with Intel 82801G (ICH7 Family) HDA Controller (rev 02), it was working on a fresh install upgraded and now it fails , I have asked in #ubuntu but no answers , so off topic not funny [00:58] 2.24.16 used to work not after update to 2.24.18 [00:58] Its a sony laptop , also suffering from the acpi problem [00:59] nixternal: yeah, i will if i can [01:13] Any idea when we'll get FXrc(1|2) in proposed? [01:14] fx? [01:16] rc? [01:16] Pici: Firefox. [01:16] I think the problem is the sony-pi module [01:16] Pici: release candidate 1/2 [01:16] ffm: RC1 is in proposed. [01:18] ffm: Also, someone just asked that on the sounder list, and got the reply that its undergoing testing, but may make it to -updates next week [01:18] Pici: not on my mirror... [01:18] Pici: cool. [01:19] ffm: I am not a developer, but I know that not all the repositories have proposed, I had to move to the main server to get those [01:33] I'm learning about packaging and I'm reading developer pages. Now I see a package (motion) not having 'ubuntu' in the version string in the changelog file. Is that wrong? Or is that right for some kind of packages? I asked on #ubuntu but there was no answer. [01:35] mouz: If the changelog file has the word unstable next to the version number, this means that the package was imported from Debian. [01:35] Whereever possible we use packages directly synced from Debian. [01:36] its much easier that way ;) [01:36] Is that an automated procedure then? [01:38] TheMuso / pwnguin : so if there's something wrong with the package, and I want to do something about it: contact upstream? [01:39] you know how dns works? [01:39] mouz: If the fix can be made in Debian, yes is better to do it there, or even upstream from Debian if its not specific to Debian/Ubuntu packaging. [01:40] mouz: start with ubuntu and work your way up; make sure it's not specific to ubuntu before pestering any upstreams ;) [01:41] nobody likes hearing their openssl is broke when it's just debian and those who branched from them :_ [01:41] pwnguin: I think I need to create a debian environment: I *think* it is a problem upstream, but I'm not 100% sure. [01:43] thats the challenge I run into. it seems silly to file bugs against debian without actually testing on a debian platform [01:45] pwnguin: That or knowing enough about Debian/Ubuntu differences to know if testing on Ubuntu is sufficient (it often is). [01:45] Virtual machines are easy enough to set up that testing on Debian isn't so hard. [01:46] ScottK: would you have a suggestion what virtual machine? [01:46] mostly i stick to the fun stuff -- x toys and the kernel [01:47] * ScottK hands pwnguin https://help.ubuntu.com/community/KVM [01:48] oh that sounds like oodles of fun [01:49] fortunately, neither of my machines appears to support it [01:49] ScottK: besides which, the kernel doesnt come from debian :P [01:50] If it's kernel, then it definitely doesn't go to Debian. [01:50] that reminds me, im supposed to rebuild and test something for the sdchi guy [01:51] pwnguin: https://help.ubuntu.com/community/VirtualBox is also good. [01:51] * wgrant uses a chroot in most cases. [01:51] mouz: at any rate, you've got a couple vm suggestions now [01:52] yes. thanks TheMuso, pwnguin, ScottK: I now can proceed :) [01:52] Ooh, new policy. [01:52] wgrant: Yes? [01:52] wgrant: where? [01:53] There hasn't been a big new one in a while. [01:53] ajmitch: debian-devel-announce. [01:53] ah [01:55] nothing too major [01:56] how do i test for return code in a debian/rules file? [01:56] eg: [01:56] No, looks like formalisation of lots of common practices. [01:56] http://pastebin.ubuntu.com/17048/ [01:57] the echo $? seems to return 'ooo-build/configure' [01:57] instead of the return code [01:57] because you have to escape the $ character in makefiles by doubling it [01:57] ah so $$? ? [01:57] yes [01:58] ok thanks :) [01:58] heh [01:58] so there wasnt a Homepage field before today? [01:59] Not in policy. [02:02] great that fixed the code to properly fail now :) [02:06] pwnguin: In Debian policy tends to define current practice, not lead it. [02:19] liw: ping [02:51] hi, i'm testing on some themes with usplash and with one strace shows a series of vm86old(0xb7d1cc8c) = -1 ENOSYS (Function not implemented) [02:51] what have to do usplash with vm86old? [02:55] video routines i think [02:57] if you check out the dependency chain of usplash, it goes to libx86, which is used to make "real mode" calls, probably to write to video hardware [02:57] yes.. [02:58] there's a comment in lrmi.c about it [02:59] used to syscall 113 [03:14] Anybody got an idea whats going on here? This is building in an intrepid chroot: http://www.pastebin.ca/1039152 The function where the problem lies is here: http://www.pastebin.ca/1039154 The bits/socket.h file gets included at the top of the c file in question. [03:18] TheMuso: My guess is the C compiler can't figure out the length of ucred, perhaps the header file that the typedef'd struct is defined in needs to be shifted down [03:19] StevenK: Yeah worth a shot, thanks for the idea. === asac_ is now known as asac [03:38] No good it seems. The fun part is that the file in question has code in it for at least Linux/Unix and Mingw32/Windows. [03:38] So the include can't really be moved further down, if only a few lines, which makes no difference. [03:39] As there are tons of #ifdefs etc. [03:40] TheMuso: Run the code through gcc -E (the pre-compiler) to see if you can resolve the struct by hand? [03:41] StevenK: Ok. [03:42] TheMuso: That advice might send you over here with a bat [03:43] StevenK: I don't know about that, it gives me something new to learn about at least. :) [03:47] pitti, superm1, ping re dkms [03:48] hi mdomsch [03:48] what's up? [03:48] superm1, #237358 [03:48] bug 237358 [03:48] Launchpad bug 237358 in dkms "dkms_autoinstaller does not support install with --force" [Undecided,New] https://launchpad.net/bugs/237358 [03:48] I'm confused [03:49] so i suppose it depends on how it was dkmsified [03:49] say I have linxux-2.6.24-14.xx-generic installed [03:49] and they bump the ABI [03:49] but in any case, AUTOINSTALL=yes should handle this fine [03:49] so new linux-2.6.24-15.xx-generic gets installed [03:50] why wouldn't dkms autoinstall "just work" - why would --force be needed? [03:50] and if it's _not_ an ABI bump [03:50] so -14.xx -> .14.yy [03:50] i think we need to see martin post his package here to see a little better what the situation is [03:50] unless /lib/modules/2.6.24-14-generic gets deleted then re-created [03:51] yeah, ok [03:51] * mdomsch adds a note to the bug [03:51] pitti, can you post it to a bzr branch and link it to the bug? [03:52] honestly i still haven't used any of the mkdeb framework for doing dkms packages [03:52] i still feel comfortable doing them by hand [03:52] i should start using mkdeb instead [03:53] and improve mkdeb then [03:53] I put it together based on how we did it with RPMs [03:53] not that I'm 100% happy with that - but it's worked fine for several years now [03:54] well i'm not happy with the idea of storing a tgz inside the deb for the source stuff [03:54] it seems unnecessary [03:54] the deb itself can just house usr/src/blah-version [03:54] that's mostly why i've done then by hand [03:55] that came about because of a marketing requirement (much as I hated it) to be able to have source-less packages [03:55] for a few big customers who were scared that "bad things could happen" if they even had the source [03:55] I haven't heard that concern in quite a few years now though [03:56] well sourceless packages are still possible without putting the tgz stuff in there, just take usr/src/ out of debian/install [03:57] i'll sit down with you over lunch one day and you can explain to me more why it was like that [03:57] and what's feasible to change [03:57] most of it's feasible to change :-) [03:57] especially on ubuntu [03:58] as so few packages exist that have used mkdeb yet [03:58] I'm completely open to any better way to do packaging [04:02] It would be kind of nice if the chunk of common code in dkms’ mkdeb postinst file were in a file provided by dkms and the postinst just sourced it. [04:04] yeah that's a nice idea [04:04] ion_, agreed [04:06] hm [04:06] could i make a proposal somewhere? ;p [04:06] i'm currently setting up crypted root for a friend of mine and i created a hook which can be used to decrypt key-files with gpg [04:06] seems to work pretty good though [05:07] are there any scripts run against the vanilla kernel to remove binary blobs, etc? [05:07] as debian does on there kernels? [05:08] if so, where can i find them? i'm looking to compile my own kernel [05:46] Hrm seems the dmsetup command doesn't like the -18 kernel. [05:46] Give me a sec and I'll get output. [05:49] This is what I get with sudo dmsetup info, this is with 2.6.24-18-generic on amd64. http://www.pastebin.ca/1039227 [05:50] meh I didn't have the damn driver loaded. [05:50] * TheMuso slaps himself. [05:50] * StevenK was about to ask .... [06:07] slangasek, pitti: time to move tcl8.5 and tk8.5 to main? [06:10] Good morning [06:10] doko: yes, think so [06:11] pitti: Do you plan on continuing to use guidance-backends for jockey? If so, is there a way I can talk you out of it? [06:11] ScottK: I'm happy to switch to something else; what would be better? [06:11] ScottK: (I saw your scrollback) [06:11] tseliot is writing a new one. [06:11] ScottK: right, I was going to look at that one [06:12] Fundamentally I xorg.conf parsing is on the way out. [06:12] I/I think [06:12] My goal is to kill Guidance as dead a possible in Intrepid. [06:12] Mythbuntu is the only other user and they'll follow your lead. [06:13] superm1: I'll give some more details on the bug [06:13] * ScottK notes the time in the US and decides to head for bed. [06:13] pitti, if at all possible, can you publish a branch? [06:13] it's be best to be able to experiment with the issue first hand [06:14] superm1: I don't have one ATM, but I can post my dkms.conf generation script, and the tarball [06:14] okay [06:14] thanks [06:14] superm1: odd, I thought I described the problem quite clearly [06:14] yeah but it shouldn't be happening [06:14] so something seems fishy about it [06:15] superm1: well, then dkms shouldn't provide the --force option for install at all [06:15] and just implicitly assume it, or so [06:15] well it shouldn't be needing to force though is the thing [06:15] either we need it in both places, or not at all [06:15] superm1: but how can it not? [06:15] it should be able to just "install" without the force [06:15] that's what dkms_autoinstaller is supposed to take care of [06:15] superm1: I can't avoid the improperly versioned modules [06:16] I need the kmods from my package, and they need to replace the ones from linux-generic, even without having a new modversion [06:16] right [06:16] so if you use AUTOINSTALL=yes [06:16] instead of AUTOINSTALL=smart [06:16] thus I need install --force [06:16] then they should take precendence without the force [06:16] superm1: I don't use smart [06:17] superm1: it's not a matter of precedence [06:17] superm1: it's a matter of the installation failing, since the 'version sanity check' barfs and aborts [06:17] oh hmm [06:17] once I --force it through, it works just fine [06:18] okay well post what you can here, i'll experiment when i get into the office in the morning and sort out a solution with matt [06:18] (what's "smart", BTW? it's not documented) [06:18] superm1: right, I will; thanks! [06:19] superm1: oh, for bug [06:19] bug 218955 [06:19] Launchpad bug 218955 in lirc "upgrade to hardy breaks lirc 0.8.3pre1 kernel modules" [Undecided,Fix committed] https://launchpad.net/bugs/218955 [06:19] superm1: you need sponsoring? [06:19] yes [06:19] 'kay, will do [06:19] great thanks :) [06:21] superm1: for intrepid, is there a source pkg? or is it all in bzr, I just checkout and debuild? [06:21] pitti, for intrepid grab the orig.tar.gz from lirc.org, and then debian/ is in bzr [06:22] i haven't added a get-orig-source rule because it's very rare that upstream actually does releases, it's mostly been snapshots of different varieties [06:24] hm so i thought AUTOINSTALL=smart was supported, but maybe i'm mixing something else up with it [06:24] grep smart /usr/sbin/dkms doesn't give anything, neither in the manpage [06:24] yeah that's what i was checking too [06:25] i'll have to check with the colleague of mine who was referencing it in the first place tomorrow [06:26] argh, /me radiates hate to that silly sf.net thing where you click on a .tar.gz link and it just saves a html page [06:29] asac: which of the two epy's with identical changelogs should I accept for SRU? [06:32] superm1: ok, all done [06:34] Keybuk: argh sysvinit argh; still no luck, apt refuses to upgrade over it [06:36] Keybuk: I'm around now [06:40] ogra: if you could take a look at my latest upload I'll appreciate it: http://revu.ubuntuwire.com/details.py?package=moodle [06:53] asac: ah, the bug report makes it clear [06:59] afternoon [06:59] calc, pong [07:01] Afternoon Hobbsee_. [07:03] TheMuso! [07:03] how goes it? [07:05] Hobbsee_: Not too bad. Yourself? [07:05] TheMuso: uni's almost over for the semester. \o/ [07:05] Hobbsee_: Awesome. [07:06] Hobbsee_: Yeah. This is good. Less tutorialising, more writing up. [07:06] Oh. Wait. :( [07:06] RAOF: Less students ... [07:06] jeebus, not done in june? [07:06] i guess its winter down there [07:07] StevenK: Yeah, that's actually nice. Lower average wait time for coffeee! [07:08] I seem to have shipped Hobart winter weather up to Sydney, though. [07:08] Yes, you bastard. [07:08] :-P [07:09] Well, except for the fact that the temperature is roughly double what it would be in Hobart :P [07:09] Here's an interesting question: do we want to sync nouveau from debian experimental? [07:10] Actually, now that I think of it, the answer is 'hell, no!'. [07:10] aww [07:10] * persia trusts RAOFs decision, but doesn't understand it: surely only very few would install such a thing [07:10] pwnguin: Sorry. I don't want to maintain the 3d stack, thanks :P [07:11] persia: That's what _you_ think. It's in experimental because it needs a git-snapshot of libdrm. rdepend of * :) [07:11] For values of [07:11] RAOF: And this is why I trust you for all things nouveau: you actually have the answers :) [07:12] for values of * equal to anything X or 3d related. [07:12] i didnt kno debian had gallium in experimental [07:12] RAOF: hah. [07:12] It's possible that libdrm will see a release before Intrepid freeze; if so, I'll be asking to sync nouveau. [07:12] pwnguin: It doesn't. [07:12] RAOF: don't you have to mark the exams too? [07:13] then whats this 3d stack [07:13] pwnguin: libdrm contains the kernel modules and associated library for the direct rendering manager. [07:14] which doesn't have 3d in particular [07:14] The 3d stack (or, technically, the usermode 3d DRI drivers) depend on libdrm, as does X. [07:14] right [07:14] Almost all X drivers (aka DDX) do *not* require a drm component; nouveau does, because it's entirely new and all drivers are moving towards having a drm component. [07:15] RAOF: i think we've talked about this before [07:15] Probably. :) [07:15] i can certainly understand not wanting to handle libdrm2 [07:15] as part of ubuntu [07:15] just thought it was wierd that you called it the 3d stack [07:15] It's the very bottom of the 3d stack :) [07:16] It also happens to not be entirely about 3d. [07:16] computer stacks grow down ;) [07:17] depends on the architecture. [07:17] actually, the next time someone decides to tell me that counting from 0 makes sense in C, i'll ask em to draw a stack [07:18] take that "its more intuitive because its how the hardware works" [07:19] Arrays are a lie in C anyway. [07:19] oh? [07:19] array[foo] <=> *(array + foo) [07:20] yes... [07:20] im not sure thats a "lie" [07:20] It is. There is no "Array". There's just a bunch of bytes, and a pointer to the bottom. [07:20] Or top. [07:20] Or, anywhere, really. [07:21] also array[foo] == foo[array] [07:21] so what exactly is an array, do you think? [07:21] * StevenK is reminded of "There is no spoon" [07:21] Mithrandir: you'd have to have a pretty lenient typing to pull that off [07:22] I'd say an array is an object that contains a fixed number of indexable subobjects. [07:22] This makes it impossible for C to have an array, except by convention. [07:22] it's all convention [07:23] Only in C. [07:23] just some conventions are enforced by runtime and compilers ;) [07:24] Right. There is no gravity, it's just a convention the universe enforces :P [07:24] well, computer programmers are designers of systems; in a sense, we're determining whether gravity exists or not [07:24] That's probably enough semantics for now. Is python-gnome2-extras installable yet? [07:25] Or should I actually ask to have various things deliberately rebuilt, rather than waiting for new versions to provoke a rebuild. [07:26] pwnguin: http://rafb.net/p/bKyL9r10.html compiles fine without warnings with gcc -Wall [07:26] RAOF: what about the kernel modules, those should be updated too? [07:30] mvo: CRACK ATTACK! https://wiki.ubuntu.com/IntrepidImprovedLinuxMeta [07:34] * wgrant wonders why we are letting X servers loose. [07:36] good morning pitti [07:36] * mvo read [07:36] wgrant: chain them up! [07:36] mvo: no, you can't play crack attack. [07:36] tjaalton: Updated in what way? The libdrm-snapshot in experimental provides a drm-modules-source package for module-assistant. === nijaba` is now known as nijaba [07:37] RAOF: oh, ok [07:38] For Ubuntu I suppose I could look into DKMS, but it's not really a priority. [07:38] right, and if 2.3.1 is enough, we should be able to pull the needed bits in the kernel [07:39] unless the nouveau guys are constantly changing/breaking things ;) [07:40] I think it's been a while since the last drm/DDX incompatibility. [07:40] that's good [07:40] if they werent breaking things, wouldnt it be time to put drm into mainline kernels/ [07:41] No; they're waiting for the great GPU memory manager wars to decay enough to have safe levels of radiation I think. [07:41] pwnguin: it already is, but the stable version is 18 months old [07:41] Mithrandir: does that snipped work out on a 64bit platform? [07:41] good morning [07:42] morning, dj holbach ;) [07:42] hiya tjaalton :) [07:42] If worst came to the worst, we could posibly push the nouveau drm module into linux-ubuntu-modules. [07:43] RAOF: I still have some optimism left ;) [07:46] tjaalton, hi there [07:46] hey dholbach [07:46] RAOF: how closely have you been following The Great Memomy Manager Wars? [07:46] liw: howdy [07:46] hey mv [07:46] mvo :) [07:47] tjaalton: Not very much. I basically trawled the archives a week or two ago for the burning remains. [07:47] RAOF: I mean, it's like redhat/tungsten are doing a lot of ttm stuff lately.. [07:48] and isn't jesse barnes an intel guy? he's been doing that too [07:48] so it's "only" anholt and keithp doing GEM [07:48] That's not an uninfluential "only", though. [07:49] mvo: hm, the removal of aptitude, libept0, and tasksel in intrepid (in favor of new apt) is your bug, not related to my sysvutils mess, right? [07:49] RAOF: hence the quotes ;) [07:49] Where's that list gone, again? I feel the need to catch up ;) [07:49] well, if intel can't decide what they want [07:50] dri-devel [07:50] Aaaah, of course. [07:50] the thread died a while ago [07:50] do you really want to place a bet? [07:50] apparently r6xx docs are about to be released [07:52] Mithrandir: well, you've convinced me. there are no arrays in C. just collective delusions [07:53] pitti: possible, I need to check, I think I didn't break the abi, I will investigate [08:10] <\sh> moins [08:11] soren: when you merge dnsmasq, can you please stop using "multiuser" for update-rc.d? It's deprecated; please use sth. like "update-rc.d dnsmasq start 20 2 3 4 5 . stop 80 1"; this change can then be sent to Debian, too [08:11] soren: that's our only change (LSB init script is in Debian now), so if they adopt it, we can sync again [08:11] pwnguin: 64-bit> yes, that's all I have those days. [08:13] soren: same for rsync [08:17] pitti: you like the game 'crack attack'? ;) [08:17] siretart: it's all StevenK's fault [08:18] <-- pwns at crack-attack [08:18] woohoo - crack-attack! [08:18] so what are the high-scores in here? :) [08:18] i had a high score of 2500 or so. donno if that's subject to inflation [08:19] 7000 or so ... [08:19] jml pwns me at crack attack. [08:19] holy cow! [08:20] it's all in the first few minutes [08:20] get a huge combo at the start, and you can chain it while garbage transforms [08:21] * dholbach realises that he's not going to win this contest and goes back to work ;-) [08:25] pwnguin: When my machine stops running something CPU intensive, I'll play you :-P [08:25] heh [08:25] 7000 seems unreal [08:26] Like most high scores, it can't be easily replicated [08:27] still, its like 20 standard deviations away from my average [08:27] What's your average? [08:27] somewhere in the 2000's [08:27] * StevenK mostly plays the Medium AI [08:28] huh [08:28] i usually play none [08:29] pwnguin: Yeah. None means no AI. Try Medium [08:29] thats not a score [08:29] Nope, since it simulates a network game [08:31] pitti: sorrry. cant you see the which was uploaded later ... otherwise the one with the smaller sized diff.gz :) [08:31] but isnt that important. the first one just contain an ununsed patch ;) [08:31] StevenK: if you want a real challenge i have a set of colorfilters you can try ;) [08:32] hey asac, thanks for fixing epiphany-browser ;*) [08:32] pwnguin: Try None set to Xtreme :-P [08:33] seb128: yeah ... another night thing. let me know if download is now as fixed as you want it ;) [08:34] pitti: did you accept midbrowser? [08:34] thanks [08:37] pitti: ok. have updated mails to be sure ... please let midbrowser in ;) [08:38] asac: I got the right one (with the smaller patch), yes [08:38] asac: mid> not yet, will do in a bit [08:40] gracias [08:48] pitti: Was update-rc.d's multiuser option's deprecation announced somewhere? I seem to have missed it :( [08:49] soren: I just sent a mail to u-d-a@ [08:50] soren: I thought it would just affect 5 packages, but it's some more (26), thus I wrote an announcement now [08:53] dholbach: you beat my hy score. [08:53] if that makes you feel any better. [08:55] pitti: Ah, ok. The TearDown wiki page should probably be changed, too. [08:55] soren: right, I'll do that once the announcement hits the ML archive [08:56] * soren hugs pitti [08:57] dholbach: (the latency in reply is due to the failing of the connection via wet string) [08:57] argh, my mail queue is stuck [08:57] piware.de [213.9.79.162] 25 (smtp) : No route to host [08:57] WTF?? [08:57] $ nc 213.9.79.162 25 [08:57] 220 box79162.elkhouse.de ESMTP Postfix (Debian/GNU) [08:58] wfm [08:58] right, and it works for other ports, too [08:58] *#$(#$* ISP [08:58] pitti: come to the land of wet string! [08:58] Hobbsee_: StevenK seems to be the master of crack-attack :) [08:58] (90% packet loss here for you!) [08:59] dholbach: yes, seems so. then i should never play him. [08:59] pitti: Your ISP is now blocking outgoing port 25? [08:59] * Hobbsee_ is not so much a fan of getting belted into the ground. [09:00] there they go [09:04] <\sh> pitti, you need a reliable mail provider? ;) [09:04] \sh: my server is fine [09:04] \sh: I need a reliable ISP [09:05] <\sh> pitti, who is it for you? :) [09:08] \sh: http://www.fbn-dd.de/ [09:10] <\sh> pitti, use hosteurope or hetzner or someone else...those isp you can kick if they do something really stupid :) [09:10] -ENODSL [09:10] soren: changed Teardown wiki now [09:11] \sh: again: I can't get DSL here [09:11] \sh: two years ago they built a shiny VDSL node in the vicinity (to handle the fiber lines here), but now there are no ports left [09:13] <\sh> pitti, grmpf...mv pitti /dev/bw/au+rhein ; install -m 755 /usr/bin/cable /home/pitti/House ; convert pitti_sadness pitti_happyness [09:14]