=== robbiew is now known as robbiew-afk === robbiew-afk is now known as robbiew_ [00:49] Hello anyone familiar with Edubuntu seed inclusion process? [00:49] I want to get 3 packages into it =) === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [00:58] hi, is anybody experiencing this: http://pastebin.com/d275a5d95 when trying to package something user pbuilder? [01:00] ockham: looks like you've broken /bin/sh in your chroot. Make it a symlink to dash [01:00] (or maybe dash doesn't handle downgrades properly, I don't know for sure) [01:00] "unable to execute installed post-installation script: No such file or directory" means in this case that it can't find the #! interpreter, which is /bin/sh here [01:00] for /var/lib/dpkg/info/dash.postinst [01:05] cjwatson: thanks for the hints. this should actually be a pretty clean lucid pbuilder chroot, so i'm a bit afraid of having to tweak it manually [01:06] is there any way to cleanly re-create the chroot? [01:06] it looks like dash's prerm is kinda "oops". === ArneGoet1e is now known as ArneGoetje [01:14] where was the announcment of the Developer Memberhip Board voting results send? [01:15] xnox: u-d-a and it was posted on teh fridge [01:17] micahg: aha thanks =) stragily I'm not subscribed to u-d-a.... [01:17] Anyhow when is the next meeting? [01:17] the helpful 15UTC every two weeks is quite vague.... =) [01:18] feb 2 [01:18] Ok thanks [01:18] xnox: if you're ever wondering when the meetings are, check the fridge calendat [01:18] *calendar [01:19] Ok I will in the feature. Thank you [02:58] james_w: What's happened to all the debian branches for source package subversion? They are apparently deleted from launchpad?! === asac_ is now known as asac === micahg1 is now known as micahg [06:57] In the new merges page [06:57] there is bugs column with bugs assigned [06:57] how do i edit it? or link bugs there? [07:02] The script looks up merge/sync bugs on Launchpad automatically. [07:06] Kk cool =) [07:06] so it will pick mine up next time it scans ;-) [07:11] good morning [07:15] dholbach: 0/ [07:16] hi xnox [07:16] =) hope your having nice morning ;-) [07:18] yeah... it will be even better when I've had my first cup of coffee :-) [07:18] thanks xnox :) [07:19] * xnox is a bit too far to make dholbach coffee [07:20] no worries :) === Guest90451 is now known as Zic [07:42] dholbach: thank you for ACKed on my syncs ;-) \0/ [07:44] :-) [07:51] are there any subtle changes in cdbs between jaunty and karmic? [07:51] a python-distutils.mk package i was able to build while running jaunty, will not build in karmic [07:51] it doesn't even attempt to run setup.py [07:52] my rules file: http://paste.pocoo.org/show/167874/ [07:53] it builds a package, which contains nothing except the entries in debian/dirs plus copyright, changelog. [08:18] Good morning [09:06] pitti: any ideas as to what causes this error? dbus.exceptions.DBusException: org.freedesktop.PolicyKit1.Error.Failed: Error parsing subject struct [09:07] tseliot: yes, in fact I do [09:07] tseliot: weird, I fixed that in early lucid already [09:07] what is it? [09:07] I'm still getting it here [09:07] a missing parameter in the self.polkit.CheckAuthorization() call [09:07] the karmic version didn't supply start-time [09:07] but it's required with lucid's policykit [09:08] oh, so I guess I'll have to fix screen-resolution-extra [09:08] tseliot: http://bazaar.launchpad.net/%7Ejockey-hackers/jockey/trunk/revision/583 [09:08] tseliot: ah; right [09:08] I thought it was from jockey [09:08] tseliot: above patch should pretty much work for s-r-e, too [09:09] pitti: ok, thanks a lot === njpatel_ is now known as njpatel [09:46] figured it out. it didn't like that my package was named -dev [10:03] has someone any idea what happen there? http://launchpadlibrarian.net/38059356/buildlog_ubuntu-lucid-sparc.ncmpc_0.16-1_FAILEDTOBUILD.txt.gz [10:04] I've seen this in several FTBFS build logs from 18 Jan and 19 Jan [10:16] ev: hi, did anything change in recent ubiquity (between alpha-2 and now) that could explain why mat_t sees bug #510033 and I don't see it? [10:16] Launchpad bug 510033 in ubiquity "It's not possible to install mp3 codecs on a fresh install" [Medium,In progress] https://launchpad.net/bugs/510033 [10:16] ev: it looks like for him no apt-get update was run in-target during the install, but for me that worked fine (with a daily) [10:18] * ev digs [10:19] ev: there was a similar issue in karmic when on a fresh install universe package lists were not available, this is why even wrote a small patch :/ [10:23] mvo: I can test it with daily if that helps [10:23] mat_t: that would be nice. I will still add code to gnome-codec-install for the case when the system is installed without network and later gets connectivity [10:24] mvo: yes, definitely [10:24] mvo: thanks a lot for looking at it! [10:24] mvo, could be that he was not online during install? [10:25] seb128: yes, I wasn't [10:25] that would do it then [10:26] mat_t: oh? you wrote "Note: the same error occurs regardless of whether internet connection is established or not" [10:26] mvo: yes, but I did not account for the installation process [10:26] aha, ok [10:27] ev: sorry for the noise then [10:27] I only referred to what happened post-installation [10:27] no worries [10:27] that explains it [10:27] :) [10:28] mvo: so will your patch cater for this scenario? [10:29] yes [10:30] mvo: brilliant [10:31] seb128: should bug 508297 rather be considered a gtk+2.0 regression ? Should I open a gtk task ? [10:31] Launchpad bug 508297 in xchat "[lucid] channels do not change color anymore" [High,Confirmed] https://launchpad.net/bugs/508297 [10:31] * xnox has sent application to become Ubuntu Contributing Developer [10:34] ttx, reassign to gtk rather than opening a new task [10:35] seb128: done [11:22] ev: what is the issue with Ensure that Jockey's apt-cache is up-to-date after install ? (just curious) === njpatel_ is now known as njpatel === dholbach_ is now known as dholbach [11:26] If you're referring to the switch of blueprints, it only becomes an issue when we have a jockey page in ubiquity. The issue itself is that the jockey backend caches apt-cache lookups, so if you run apt-get update (via "update this installer", for example) while the backend is running, and you get to the jockey page, it wont be using the updated apt cache. [11:27] it's not a major deal, as the backend goes away after 10 minutes [11:27] mvo: ^ [12:28] ev: aha, thanks. I was just curious if its some limitation in python-apt or apt [12:28] nope :) [12:45] I wonder... if I drop a karmic libc6 on a hardy system (well, chroot), do things blow up? [12:47] very likely [12:49] ttx, kirkland: do you recall what the issue was in hardy that caused all instances of scsi-modules-$version-di in the archive to be pulled into the server CD, and is that resolved in post-hardy versions? [12:50] (it's prompting me to do some needful NBS pruning of kernel binaries in hardy-{proposed,updates,security} which we don't have an automatic way of taking care of - but it would be good if we didn't have to worry about this before lucid point releases) [12:50] slangasek: that doesn't ring any bell [12:50] ok [12:50] directhex: that was kind of my thinking [12:50] I may just do intrepid/jaunty vms as interim steps in my search for the bug [12:51] lamont, i had enough explosions with sid libc on non-sid debian to not trust it [12:51] slangasek: we should now be able to get rid of cglib2.1, btw [12:51] lamont: I'm not aware of any specific reasons it would explode [12:51] directhex: ah - very good data point, ta [12:51] ttx: woohoo! :) [12:51] slangasek: eucalyptus now build/runs from libcglib-java [12:51] (2.2) [12:51] slangasek: coolness. [12:52] lamont: I may just be exposing my own ignorance, though :) [12:52] slangasek: there are a few universe packages that still depend on libcglib2.1-java but libcglib-java provides it. [12:53] slangasek: let me know if there is anything more I should do on that subject [12:53] slangasek: you're the one who encourages boy scouts to throw gasoline on the fire, aren't you? [12:53] on debian it's usually locales that go to poop when you update libc... but ubuntu's locales are handled differently [12:53] not sure how automatic those removals are [12:54] ttx: semi-automatic [12:54] Question about main: when a source package gets promoted into main, do all the resulting binaries also end up in main ? Or do they each need to be pulled by something ? i.e. can a main source package have main and universe binaries ? [12:54] yes, a main source package can have both main and universe binaries [12:55] slangasek: ok, cool. [12:55] and if you find a universe source with main packages, then you have found a bug. or warty. [12:55] lamont: heh [12:56] That can still happen if a binary starts being provided by a different source. [12:56] 'twas one of those lolwut? moments [12:56] It happened for a bit recently with the switch to distribute from python-setuptools source. [12:58] ScottK: that should show up on components-mismatches, though? Or was this "old package forgot to say that it no longer owned it"? [13:41] mvo: ping === marjo_ is now known as marjo [13:56] cjwatson: currently we don't pass the vga= parameter to vga16fb, right? Any ideas as to how bug #509328 can be fixed without this parameter? [13:56] Launchpad bug 509328 in nvidia-graphics-drivers "Lucid Alpha2: Plymouth does not work with the current nvidia driver" [Medium,Confirmed] https://launchpad.net/bugs/509328 [13:56] tseliot: that's correct, because using vesafb breaks suspend/resume [13:56] tseliot: I believe the plan of record is to fall back to vga16fb, not vesafb [13:58] cjwatson: right but currently vga16fb doesn't seem to allow plymouth to work. Is there some parameter we can pass vga16fb? [13:58] * tseliot wasn't suggesting that we use vesafb [13:58] pass it to do what? [13:59] tseliot: that I don't know, Keybuk probably would [13:59] cjwatson: ok, I'll send him an emai. Thanks [13:59] I assume that plymouth explicitly rejects it or something at the moment, since vga16fb should present a framebuffer much like anything else [14:00] or that we aren't arranging for vga16fb to be loaded if nothing better is available [14:00] ok [14:02] from bug #506717 and an examination of the initramfs scripts, I believe vga16fb should be getting loaded correctly in the initramfs [14:03] Launchpad bug 506717 in plymouth "[Lucid] plymouth does not display when using nvidia drivers" [High,New] https://launchpad.net/bugs/506717 [14:03] i.e., the affected users have it loaded /after/ boot, and the only thing that loads it is the alias, so it should load just as well from the initramfs [14:05] right [14:26] Check out http://bzr.debian.org/bzr/users/xnox-guest/merges/ [14:26] =) [14:26] It's a webpage =) [14:27] Revamped new merges page [14:27] a little bit =) [14:27] http://bzr.debian.org/bzr/users/xnox-guest/merges/index.html === dendrobates is now known as dendro-afk === dmart_ is now known as dmart === dholbach_ is now known as dholbach === highvolt1ge is now known as highvoltage === tkamppeter_ is now known as tkamppeter [15:02] kees, you have a minute? i've a questoin about ssh known hosts format === dendro-afk is now known as dendrobates === yofel_ is now known as yofel === herb__ is now known as herb === beuno is now known as beuno-lunch === BenC1 is now known as BenC [16:58] smoser: hello! I'm here now, going through IRC backlogs. :) [16:59] kees, i bothered jdstrand , he answered. i was somewhat worried that in my known hosts file the 3rd field always started with same 28 chars or so [17:00] it appears to be key header, which always base64 encodes the same, so not an issue (at least i dont think so) [17:01] lool: ping. Do you have a minute? [17:02] smoser: ah-ha, yes. that's correct (key header) [17:03] i was afraid i was getting little or no randomness in key generation [17:06] Laibsch: Depends how long the minute ;) [17:17] doko: did the build work? [17:24] ccheney: the build takes about 40h ... [17:29] doko: oh so won't be done until tomorrow i suppose? [17:30] doko: was the problem in the past it failed to build or the resulting binaries just failed to work? if failed to build it sounds like it still going is a good sign :) === beuno-lunch is now known as beuno [18:01] lool: ping. I sent you one more question. If you another minute... [18:08] bryce_, pitti, seb128: today's updates removed xserver-xorg-video-intel on an intel-GPU based system... any idea why? [18:26] Is there a way to speed up key processing of plymoth in lucid? Entering the passphrase for my disks is really annoying now. [18:33] slangasek: hi - re bug 507778 [18:33] Launchpad bug 507778 in acpid "Please merge acpid (1:2.0.0-1) from Debian testing" [Undecided,New] https://launchpad.net/bugs/507778 [18:33] slangasek: does it make sense to pull 2.0.0 in lucid LTS? [18:43] mathiaz: I haven't looked at what's different in 2.0.0 [18:44] mathiaz: though I do know there are some changes in the Debian package that will need to be gone over carefully in a merge [18:45] slangasek: the debian changelog entry says: New Upstream version from new source tree that already incorporates the [18:45] slangasek: netlink patch.. [18:45] slangasek: I'm not sure how stable is this *new* tree with regards to LTS [18:48] slangasek: AFAICT there isn't any major packaging change - kacpimon is a new binary package that has been added [18:50] yes, I don't know either [18:51] kees: jdstrand: mdeslaur: what's your opinion on bug 510732? [18:51] Launchpad bug 510732 in openssh "OpenSSH server sshd_config PermitRootLogin -> NO" [Wishlist,Incomplete] https://launchpad.net/bugs/510732 [18:51] mathiaz: we defer to cjwatson. I would support the change, though. [18:52] kees: could you add a comment to the bug? [18:54] cjwatson: so, what do you think about this ^^ ? It's been a long-standing request, and I kind of like it. [18:54] mathiaz: what do you think of it? [18:54] mathiaz: I'm with kees. I too like it, but realize it cuts down on usability [18:54] mathiaz: the logic has been that rootlogin isn't possible by default because there is no root password. [18:55] kees: yes - that was my first answer [18:55] people argue that if they enable the root account, they want to use it [18:55] jdstrand: and it's true that the choice to be made is between usability and multiple layer of security [18:55] right [18:56] that said, it is something I typically change on machines I administer (and if I can get away with it, AllowUsers) [18:56] * kees ♥ AllowUsers [18:56] (not suggesting we try to do anything with AllowUsers) [18:57] *totally* [18:57] so, here's a thought... [18:57] we have an implicit possible with gdm to disallow the root user to log in, even if they have a password set. [18:57] perhaps we should duplicate this with SSH? [18:58] kees: an implicit possible with gdm to disallow the root user? [18:58] interesting thought, though there is a similar discussion with sudo and gksudo [18:58] kees: what do you mean by that? [18:58] mathiaz: oh, er, total typo. s/possible/policy/ [18:58] Aren't there system risks with running Gnome/KDE as root? [18:58] it is interesting in this case that the gui app is more restrictive [18:58] ScottK: yes, very much [18:58] * ScottK wouldn't see blocking root login as primarily a security issue [18:58] It's a this won't work issue [18:59] ScottK: right, that's why gdm blocks root. [18:59] kees: right. there seem to be a policy mismatch here [18:59] * slangasek runs GNOME as gdm [18:59] it is a known account which makes it particularly easy to brute force (and everyone wants it) [18:59] kees: that being said, openssh is not installed by default on desktops [18:59] but I'm saying that it might make sense to carry that logic to SSH [18:59] So it's tangential to the security issue [18:59] slangasek: so did I briefly :) [18:59] kees: what the reason for disabling root login in gdm? [18:59] ScottK: I don't think it is tangential [19:00] kees: it seems that the target users are different here [19:00] There's no system reliabilty reason to avoid ssh as root. === elmo_ is now known as elmo [19:00] ScottK: if you have an enabled root account *and* allow root passwords (also the default), then it makes it possible to brute force [19:00] kees: gdm policy is to disallow end users to shoot themselves in the foot [19:00] mathiaz: AIUI, lots of system files can get trashed. gnome/kde don't support it, etc. [19:00] jdstrand: Certainly, but that's not why gdm doesn't allow root loging [19:00] login [19:00] kees: are users installing openssh more literate? [19:01] ScottK: oh, I thought we were talking about ssh still [19:01] kees: ie they know more about the root account [19:01] kees: and when they enable the root account they know what they're doing? [19:01] mathiaz: I would say a greater percentage of people installing openssh-server understand the root account [19:01] running the whole X/Gnome stack is a lot of code that wasn't really written with running as root in mind [19:01] mathiaz: and I don't think that people enabling the root account know what they're doing always, which is why I think there's value in making the SSH default be tighter [19:02] again, I'm for it, but defer to cjwatson [19:03] kees: the other argument for defaulting PermitRootLogin to no is that's what upstream recommends in their documentation [19:03] (even though the default is yes) [19:04] mathiaz: yeah, agreed. mostly we would need to convince cjwatson. :) [19:04] Sounds like upstream should change then [19:11] mathiaz: http://cheezburger.com/View.aspx?aid=3094191616 [19:12] kees: lol [19:13] kees: that's *very* helpful in taking a good decision :) [19:13] mathiaz: yeah, I've been using that site for diagrams a lot lately, super easy [19:13] kees: the whole question is actually about the *size* of the different circles [19:14] kees: if blue > brown, then default PermitRootLogin to yes [19:14] kees: if blue < brown, then default PermitRootLogin to no [19:14] mathiaz: exactly. [19:15] mathiaz: actually, maybe not. [19:15] kees: AFAICT the *current* representation relies on the number of letters in the circle label [19:15] PermitRootLogin to no protects the dark green area. [19:16] kees: minus the blue area [19:16] mathiaz: yup. comment posted to the bug detailing this logic. [19:17] kees: awesome! [19:18] kees: jdstrand: thanks for taking the time to look at this [19:19] mathiaz: sure, no problem. it's come up before, but the brute-forcing attempts have been getting stronger lately, so it's valuable to revisit it. [19:19] slangasek: i don't recall that at all, sorry [19:22] mathiaz: np! :) [19:34] NCommander: Hi, I was told by charlie-tca that you were the best person to speak to about helping to build packages for the Ubuntu PS3 port (PowerPC) [19:35] I want to contribute === gnomefreak76 is now known as gnomefreak === mrpouit is now known as mr_pouit === ajmitch_ is now known as ajmitch === arand__ is now known as arand [20:22] Wow this is busy === cjohnston is now known as notcjohnston === notcjohnston is now known as cjohnston [20:37] Laibsch: It's ok, don't ask to ask, just leave me your questions and I should get to them [20:37] lool: I don't ask to ask, but I was under the impression that a ping here grabs your attention while simply writing to you in an ongoing chat doesn't [20:38] just wanted to make sure our discussion isn't lost in the void [20:53] hey, anyone available to sponsor a merge? I filed the proposal about 4 days ago. I'm not in any particular hurry, just don't want it to rot: https://code.edge.launchpad.net/~statik/ubuntu/lucid/protobuf/merge-bug502654/+merge/17571 [20:54] Laibsch: it's actually exactly the same, except I get two red channels instead of one :-) [20:55] lool: OK, thanks [20:58] statik: At this point in the cycle, that's not so long. [20:59] ScottK: i can only imagine. i'm buried in work myself and i don't have to deal with sponsoring :) [21:26] where does sound-juicer get track names from? (and more to the point, when it decides that the musicbrains answer is available, how do I tell it "no, that's wrong"? [21:26] and yeah, echan and all that [21:30] siretart: Debian seems to have fix for the gst-plugins-bad0.10 FTBFS. [21:35] lamont: musicbrainz for one [21:36] lamont: And very likely cddb/freedb. [21:39] TheMuso: well, it's telling me that musicbrains doesnt have it, so I expect "not that" [21:39] more to the point, sometimes it's getting it very right, other times wildly wrong - for the same CD set... [21:39] ergo, I want a way to influence it's voting... [21:40] lamont: heh right. [21:40] * TheMuso has found that the shell script abcde is the most flexible method of ripping for his needs. [21:42] does it get track info? [21:43] actually, is there any way to have that data encoded on the CD? or is it all magic lookups based on hashes? [21:44] lamont: Yes it gets track info, I think it may have some support for musicbrainz, but it mostly uses freedb. Note that there is no GUI for abcde. [21:44] lamont: I don't think this can retrieve cd-text. [21:49] ok [21:49] thanks [21:51] is there any reason why fuse-utils provides a /sbin/mount.fuse helper but no umount.fuse helper? [21:52] kees: convince upstream. I'm not going to deviate from them [21:55] kees: and frankly, I'm fed up with having the discussion every ten minutes [21:55] (this isn't getting at you, but at the people who won't read) [23:35] hi, is there any easy way to rename a manpage when installing it using debian/.manpages? === robbiew is now known as robbiew_