[12:04] yes, there's a known bug about vmware disks IIRC [12:08] which is a trivial fix === r0bby [n=wakawaka@guifications/user/r0bby] has joined #ubuntu-devel [12:09] s/fix/to fix/ [12:09] s/a // ;) === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-devel === jcape [i=jcape@71.194.176.102] has joined #ubuntu-devel [12:18] Curious: /etc/sudoers has $admin in it by default. [12:18] %admin I mean. [12:18] o mysql-client-4.1 mysql-server-4.1 {mysql-dfsg-4.1} [12:18] infinity: ok to demote to universe? [12:18] admin is uid 1000, there is also a group admin with gid 1000. [12:18] This was the first group I created during install. [12:18] wasabi_: admin's only a uid if the initial user you create is called 'admin' [12:18] which is a bit of a corner case :) [12:18] Ahhhh. [12:18] I was wondering if %bob would be in there in the case of that. [12:18] I didn't expect anyone to do that ... [12:19] % denotes a group [12:19] My first (and only) user is always admin. [12:19] hmm, that's pretty awkward to work around [12:19] adm seems more appropiate. [12:19] I think I might just ignore it, it produces essentially the same result in your case anyway [12:19] %adm in there. [12:19] no, adm is for viewing log files, it just has a crap name === ogra imagines that former win admins might call the first user admin [12:19] Ahh. [12:19] Well hell. [12:20] we went through all the possibilities when we introduced the admin group, and none of the existing groups seemed suitable [12:20] I'd be all right with establishing a default admin group in a system gid range. [12:20] there is one [12:20] You introduced the group? [12:20] oh I just create the admin user. [12:20] What's the gid supposed to be? [12:20] it just gets created after user creation [12:20] adduser --system, no fixed gid [12:20] Hmm. So mine is 1000 because my user was named admin [12:20] if an admin group already exists as a result of initial user creation, the system group won't get created [12:20] Ahh. [12:20] I see. That works out. [12:20] I think that's better, it fails more gracefully in your case [12:21] Yeah. === lotusleaf [n=sigiloff@unaffiliated/lotusleaf] has joined #ubuntu-devel [12:21] I am now clear! [12:21] in general we try to use adduser rather than fixed ids [12:21] Yeah, I follow. [12:21] I wish there was some way to prevent clashes in distributed user bases. [12:21] MS figured it out. [12:21] theoretically most of the existing fixed ids can be phased out, but I haven't had the stomach for the necessary combo base-passwd / adduser hacking yet [12:22] there's probably a good wine for that [12:23] At some point in the future can we stop with this too: admin:x:1000:1000:Administrator,,,:/home/admin:/bin/bash [12:23] ,,, [12:23] I get asked silly questions from strangers in Tescos ... "which wine is good for chicken?" "which wine is good for beef?" "which wine is good for hacking on console-tools?" etc. [12:23] Haha. [12:24] wasabi_: er, you do know what gecos is, right? [12:24] Nope. ;) [12:26] wasabi_: try chfn [12:26] Oh. Heh. [12:27] the fifth field of passwd is the gecos field and contains several comma-separated components [12:27] Treenaks: brief description of promotion/demotion and main inclusion reviews now in DeveloperResources === ssam [n=ssam@88-104-137-90.dynamic.dsl.as9105.com] has joined #ubuntu-devel === lotusleaf [n=sigiloff@unaffiliated/lotusleaf] has left #ubuntu-devel ["melting"] [12:29] Is it just me or is /var/run/nscd missing by default? [12:29] Nope, not just me. === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === mpt [n=mpt@219-89-144-79.jetstart.xtra.co.nz] has joined #ubuntu-devel === LaserJock [n=mantha@ubuntu/member/laserjock] has joined #ubuntu-devel === thierry_ [n=thierry@modemcable082.90-81-70.mc.videotron.ca] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === schlomo [n=chatzill@ASt-Lambert-153-1-50-98.w83-204.abo.wanadoo.fr] has joined #ubuntu-devel [12:52] Hi sorry to bother you [12:53] it's possible to load fglrx in 2.6.12-9-386 breezy ? *ERROR* Unable to the open some already present DRM kernel module! when load fglrx with modprobe [12:54] I have to update kernel ? it's a known problem ? [12:54] it's not possible sorry [12:56] didn't find anything on bugzilla === Alinux [n=Ubuntu@p54A39848.dip0.t-ipconnect.de] has joined #ubuntu-devel === lfittl_ [n=lfittl@83-65-242-100.dynamic.xdsl-line.inode.at] has joined #ubuntu-devel === seb128_ [n=seb128@ANancy-151-1-32-248.w83-196.abo.wanadoo.fr] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === womble [n=mpalmer@sponge.solutionsfirst.com.au] has joined #ubuntu-devel === sfeehan [n=sfeehan@pool-64-223-118-4.burl.east.verizon.net] has joined #ubuntu-devel [01:23] Kamion: Yessir. Demote away. [01:27] infinity: done === TerminX [n=terminx@64.62.190.212] has joined #ubuntu-devel === seb128 [n=seb128@ubuntu/member/seb128] has joined #ubuntu-devel [01:34] Oh crap. [01:34] infinity: Look at jnethack on amd64! [01:35] infinity: It's been reported as a Debian bug, too. [01:36] StevenK: Don't wanna! :) [01:36] StevenK: (Will look at it in just a bit, after I've settled into my morning routine) [01:36] mkfontdir segfaults [01:36] Who has access to keyring RT? I've sent an e-mail to it, and the reply quoted my e-mail saying "This has no content" [01:37] daniels: It seems that bdftopcf is throwing out crap mkfontdir can't understand, from my (limited) investigation. [01:37] score === swat_ [n=swat@82-45-238-129.cable.ubr02.enfi.blueyonder.co.uk] has joined #ubuntu-devel [01:39] hi guys, anyone about who knows about gstreamer/rhythmbox. i'm trying to figure out if i am experiencing a bug, or just something gone bad on my machine === zul [n=chuck@ubuntu/member/zul] has joined #ubuntu-devel [01:39] swat_: cvs snapshot that dapper uses isn't fully ported yet. === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [01:40] HiddenWolf: i see - it could be a gstreamer problem though. basically if i tell it to play using my internal card on my laptop the standard intel one, all is fine. if i give it the usb creative card, it bombs out and goes mental [01:40] if this is to be expected at this stage of development fair enough, i was just wondering if it was worth filing a bug report for [01:41] swat_: please do. [01:41] it wasn't a problem in hoary, so i assume it is some sort of issue - rather than not being supported [01:43] swat_: new kernel, hotplug to udev transition, there has been messing with the default sink, and the gstreamer 0.8 -> 0.10 thing. It can be a lot of things. [01:43] so i should submit a bug to the rhythmbox thing anyway [01:43] and it'll get bounced around one assumes [01:44] right === marcin [n=user@194.114.146.58] has joined #ubuntu-devel === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-devel [01:55] elmo: can you please remove evolution-caldav from Dapper? === rikai [n=gtk2@pool-70-105-231-88.port.east.verizon.net] has joined #ubuntu-devel [02:15] elmo: fyi, x11-common should go into main, and when it does, xorg-common and xserver-common should be melanied [02:16] daniels: k === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [02:18] debian takes legality dead-serious. :/ [02:19] Software. Daniel Carrera [23] wondered how he is [02:19] supposed to fulfil the source code requirement of the GNU [24] GPL when [02:19] he is handing out OpenOffice.org CDs during an exhibition. Andrew [02:19] Suffield [25] explained that the easiest way is to prepare copies of [02:19] the source and give them to anybody who asks for them. [02:19] odm [02:19] isn't that a bit, ehm, zealous? === PuMpErNiCkLe [n=Moitd@CPE000d8802a749-CM0012c9a9a6dc.cpe.net.cable.rogers.com] has joined #ubuntu-devel [02:21] argh [02:21] people do get asked for source code when handing out CDs at exhibitions [02:21] at least occasionally in the context of whether they're meeting licence requirements [02:21] it's a lot easier to have your answer prepared than to be stumped [02:22] Shouldn't it be enough to have your awnser "sourcecode is published at XYZ location" ? [02:23] If it isn't, than isn't shipit in violation of the gpl? === stub [n=stub@ppp-58.10.73.242.revip2.asianet.co.th] has joined #ubuntu-devel [02:23] shipit includes a written offer last I checked [02:24] (which I don't think anyone's ever taken up ...) [02:24] it does [02:24] you have to either (a) accompany binaries with source (b) include written offer (c) if non-commercial, pass on the offer of source you received [02:25] when I've handed out Debian CDs at exhibitions at the past, we went for (b) [02:25] but that does require having a contact address of somebody willing to mail out CDs [02:25] just pointing them to a web archive isn't sufficient, IIRC, because you can assume that a fair proportion of people who are wanting to get CDs from you rather than do the download themselves won't have the bandwidth to download the source either [02:27] I'm not *entirely* sure the GPL says that straight out but it seems reasonable behaviour [02:28] i thought there was some form of media restriction on the source offer [02:28] being that you have to at least make it available through the same medium [02:29] but I could be mistaken [02:29] "on a medium customarily used for software interchange" === HiddenWolf thinks floppy disks [02:29] I think you could argue convincingly that archive.ubuntu.com plus the Internet is such a medium [02:29] (e.g.) [02:30] but I'm not sure everyone actually agrees with me there, so shrug [02:30] night guys. [02:30] dude, we don't use the interweb for anything [02:30] night dholbach [02:30] the last paragraph of GPLv2 section 3 does suggest that access to copy from a designated place is not quite the same as distributing on a medium, though [02:31] so, whatever; it's easy enough to comply with the strictest reading in case anyone asks, and it makes the odd person happy and saves you aggro [02:34] imo, it should be thought of as an honor to be able to have source CDs for people [02:39] we kick out source CD images as part of our daily CD image build process [02:39] (yes, we have been asked for them, too) [02:40] daniels: nice changelog on xorg. :) [02:42] daniels: xlibs-dev description: pac -> package [02:45] (oops) [02:48] daniels: Any urge to look at the mkfontdir 64-bit hatred? === jsgotangco [n=jerome@ubuntu/member/jsgotangco] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === slomo_ [n=slomo@ubuntu/member/slomo] has joined #ubuntu-devel === wftl [n=mgagne@salmar.com] has left #ubuntu-devel ["Kopete] === Kinnison [n=dsilvers@haddenham.pepperfish.net] has joined #ubuntu-devel === ozamosi [n=nnnnnnnn@h203n8c1o1049.bredband.skanova.com] has joined #ubuntu-devel === mahangu [n=mahangu@unaffiliated/mahangu] has joined #ubuntu-devel [03:16] infinity: awakeish? === Burgundavia wonders whether to respond to sabdfl with "sorry, ubuntu-devel is not for bug reports. Please file them in LP" [03:19] tseng: Maybeish. [03:19] infinity: hardware detection on my desktop flight 3 install is pretty bunk, i seem to remember you are the guy [03:19] infinity: i'd be lying if I said yes [03:19] Burgundavia: Go for it. Please. [03:19] infinity: i'll check it out if I get time after looking at a few xorg issues === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [03:20] tseng: Keybuk is the guy, actually. [03:20] ah, well its past his bedtime [03:20] tseng: Define "bunk". [03:20] its not loading modules [03:20] in single user, my nic module is not loaded for example [03:21] in a normal boot X goes totally bonkers and displays garbage on my dvi, seemingly hung [03:21] so not much clue from that angle yet [03:21] Burgundavia: i don't think that would be inappropriate [03:22] a startx in single user says "you have no devices, jackass!" [03:22] jdub, yes, that is why I didn't do that [03:22] tseng: Hrm, S10udev is meant to catch all the devices that didn't get loaded in the initramfs. [03:22] Burgundavia: or backwards -> i think it would be appropriate [03:22] jdub: double denial is always tricky. :) [03:22] tseng: You'll definitely want to either file a bug or talk to Keybuk directly though. [03:23] infinity: indeed. [03:23] HiddenWolf, jdub, gahhh! I give up [03:23] tseng: make startx sign the code of conduct. :) [03:28] what is the copyright/license for debian/* ? [03:29] LaserJock: generally debian/rules is licenced under the gpl (look in the file itself), and the rest isn't copyrightable [03:29] daniels: what does copyrightable mean exactly? [03:30] LaserJock: if you can claim copyright or not [03:31] so does that mean public domain? [03:31] LaserJock, anything that is not copyrightable is not copyrightable. Effectively the same as PD [03:32] interesting [03:32] PD is for things which have had their copyright lapse, but they are still technically under copyright [03:32] if I write 'the grass is green', that doesn't deserve copyright protection [03:32] daniels: not? ;) [03:32] HiddenWolf: ? [03:32] daniels: joke. [03:32] daniels, that is debatable, depending on the country [03:33] basically, you need to demonstrate that you actually did something [03:33] Burgundavia: of course you can claim copyright, to get it is a different story. [03:33] but packaging is a lot of work so I'm not sure why it would be non-copyrightable [03:34] LaserJock, the descriptions are copyright, probably the same license as the program, as a derived work on the main body [03:34] scratch that [03:35] probably gpl, as part of the packaging, as daniels said [03:35] LaserJock: it's a lot of work, but it is not unique, your achievement specifically, or otherwise protectable [03:35] LaserJock: You're doing work following a guideline, the guideline might be copyrightable, the work itself not, afaik. [03:35] LaserJock, however, if you quoting a small piece of the work, US copyright law has a thing called fairuse. This is what reviewers use [03:36] hmm, it seems pretty unique to me but the docteam has gotten me all paranoid about licenses ;-) [03:36] LaserJock, you should be paranoid about copyright. It can and will bite you in the ass [03:38] It just seems odd to me that we don't have to say anything about the copyright of the actual packaging when we are packaging === ozamosi [n=nnnnnnnn@h128n1c1o1049.bredband.skanova.com] has joined #ubuntu-devel === psusi [n=phreak@54.161.205.68.cfl.res.rr.com] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === mhz [n=mhz_chil@pc-252-84-215-201.cm.vtr.net] has joined #ubuntu-devel === r0bby [n=wakawaka@guifications/user/r0bby] has joined #ubuntu-devel === Gman_ [n=gman@nwkea-socks-1.sun.com] has joined #ubuntu-devel === johanbr [n=j@d154-20-176-233.bchsia.telus.net] has joined #ubuntu-devel === r0bby_ [n=wakawaka@cpe-204-210-133-115.hvc.res.rr.com] has joined #ubuntu-devel [04:15] what is the right way to go about getting something included in the initrd on the install cd? how exactly is that built up from the udebs? === Alinux [n=Ubuntu@p54A39157.dip0.t-ipconnect.de] has joined #ubuntu-devel [04:16] psusi: For testing, just rebuild debian-installer with your udeb in build/localudebs [04:16] psusi: For the "real thing", what gets included is defined by archive priorities. === mpt [n=mpt@219-89-144-79.jetstart.xtra.co.nz] has joined #ubuntu-devel [04:17] infinity, so building debian-installer from source installs the udebs in build/localudebs to a chroot somewhere? then snatches the kernel and initrd out of that chroot? === psusi thinks it is time to udebify dmraid === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [04:21] what I'm getting at is if the udeb's postinst adds scripts to /usr/share/initramfs-tools and calls update-initramfs, that will get the right stuff into the initrd for the install cd yes? [04:26] Erm, no. I'm not positive how that all works, now that d-i has moved to initramfs. === infinity caves in and uploads a "neon24" source package for now, until stuff like openoffice and bazaar have been adapted to the new API. [04:29] any idea why impilinux.org.za redirects to ubuntulinux.org? [04:29] for that matter, why don't I see any udebs in the archive? [04:30] Burgundavia: isn't impilinux the distribution that sabdfl bought, that'll put a paid ubuntu deriviate translated to za languages to market? [04:30] I'm sure I read something about it, once. [04:30] HiddenWolf, yes, but impi should still have it own page [04:31] http://mybroadband.co.za/nephp/?m=show&id=828 [04:31] psusi: Err.. I do.. (http://archive.ubuntu.com/ubuntu/pool/main/b/binutils/ is an example) [04:32] Burgundavia: perhaps they'll put up a new site when they are ready? === PuMpErNiCkLe [n=Moitd@CPE000d8802a749-CM0012c9a9a6dc.cpe.net.cable.rogers.com] has joined #ubuntu-devel [04:32] psusi: If you mean "why don't I see them in Packages.gz", the answer to that should be reasonably obvious. udebs shouldn't be installed on running systems. [04:32] ohh... I see [04:32] HiddenWolf, no idea, but http://www.impilinux.org/ gives me a blank page [04:33] Burgundavia: well, you'd expect them to use launchpad and such, but there is nothing there either, afaik [04:33] Burgundavia: guess it's a ghost untill it's promoted. [04:34] HiddenWolf, I met the impi guys at UBZ === psusi wonders WTF Build-conflicts: means... can't be build at the same time as that package??? [04:35] surprisingly, yeah [04:35] what kind of sense does that make? [04:35] don't you build only one package at a time? [04:36] can't be INSTALLED at the same time as that package is being built [04:36] Burgundavia: *shrug* Can't imagine sabdfl sinking a lot of money into something without a good plan, but I'd have no clue as to the plan. [04:36] ohhh... wow... weird [04:36] HiddenWolf, personally I think Impi might just fail, but hmm... [04:37] 3 versions, 3 different base packages does not confidence build [04:37] 2 different DEs [04:37] elmo: Can you NEW neon24, please? (Yes, this is IncreasingDuplication, go me, it should be hammered out in the next few weeks/month) [04:37] Burgundavia: I'd expected to see those guys active in the ubuntu world, links everywhere and such, but that hasn't happened yet. Who knows. [04:38] HiddenWolf, one of the impi guys was in #ubuntu-motu toda [04:38] infinity: new for what? main or universE? [04:38] elmo: main, sorry. [04:39] psusi: "Having this package in the build chroot completely hoses MY build" is a pretty common case, actually, but one that isn't tripped on enough (due to build chroots usually starting "clean") for people to add Build-Conflicts very often. [04:39] wackey === psusi wonders if he can't just mv dmraid*.deb dmraid*.udeb [04:40] No. [04:40] psusi: udebs follow a different policy. === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [04:41] I mean for testing ;) === freeflying [n=freeflyi@61.190.65.8] has joined #ubuntu-devel [04:43] Burgundavia: *shrug* One guy, no website, big plans. Time will tell seems to be the correct saying for the matter. [04:43] HiddenWolf, I think it is two [04:45] infinity, btw... you DID say you had no idea why that dmraid bug was assigned to you, and I should reassign it to fabio right? or did I imagine that? [04:46] psusi: Yes. [04:46] good [04:49] jdub, are there trademark issues with nUbuntu? [04:52] regarding trademarks this seems to be a violation in some way. www.ubuntu.se Its in english so you can read it. Its a commersial company I think. (given that ubuntu is trademarked and it applies to swedish law - which i think it does..) [04:54] kent, remember that ubuntu also means something and that appears to have nothing to do with comoputers [04:54] kent, see also ubuntu.org === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [04:54] kent: shees, what a philosophical babble. :) [04:55] HiddenWolf, awful. All this loving stuff [04:55] Burgundavia, ah :) [04:55] Burgundavia: ahaha. :) [04:55] ahahah^2 [04:55] kent, however, nUbuntu is a derived distro [04:55] Burgundavia: the idea is good, but it's a tad much. :) [04:58] Burgundavia: read ubuntu.org [04:58] Canonical Go Ahead [04:58] Today I received an email from Canonical about using the Ubuntu name and logo. We have got permission from them to use the name "Ubuntu" and Ubuntu logo. So people having doubts about us using the Ubuntu name and logo, you can stop worrying. :) === mpt [n=mpt@222-154-180-221.jetstream.xtra.co.nz] has joined #ubuntu-devel [04:59] HiddenWolf, yes, but what are you using it for? [04:59] Burgundavia: I'm not using it. I just found out it existed 10min ago. :) [05:00] HiddenWolf, where is that quote from? [05:00] Burgundavia: nubuntu.org [05:00] ah [05:00] jdub, nevermind then [05:01] Burgundavia: google does wonders. ;) [05:01] HiddenWolf, I haven't checked their page is several days [05:01] they did hit the front of digg [05:02] hm [05:02] oh man [05:02] so what on earth is udevplug actually doing in dapper? [05:02] mkdir("/dev/.udev/queue", 0755) = -1 EEXIST (File exists) [05:02] nanosleep({0, 10000000}, NULL) = 0 [05:02] I'd rather see that energy move into ubuntu proper, but I guess it's good to have a healthy set of derivs to work with. [05:03] i see *tons* of those in strace output, until finally: [05:03] dilinger: What's it do, or what's it supposed to do? [05:03] HiddenWolf, sabdfl is right about derivs [05:03] nanosleep({0, 10000000}, 0) = ? ERESTART_RESTARTBLOCK (To be restarted)--- SIGALRM (Alarm clock) @ 0 (0) --- [05:03] exit_group(1) = ? [05:03] infinity: yes [05:04] dilinger: It's supposed to do bus scans and generate hotplug events for coldplugged hardware. [05:04] dilinger: I'm not sure if that's what it IS doing for you. :) [05:04] 30% of the way through the 2.5M strace file, it starts w/ those [05:04] that's the entire rest of it [05:05] Well, I guess you've just volunteered to talk to Kebuk about it. :) [05:05] Keybuk, too. [05:05] where is he? [05:05] in bed. [05:05] d'oh [05:05] I should hope [05:05] any idea what 's up w/ udev permissions, too? [05:05] crw-rw---- 1 root root 1, 3 2006-01-17 22:57 /dev/null [05:05] dilinger: He should be around in a few hours... [05:05] ok [05:06] dilinger: it's crack, up to you to decide if it's a good or a bad trip. :) [05:06] My /dev/null is fine. [05:06] what's really strange is that once i added the strace to udevplug, it booted faster [05:06] before, it was hanging for like 10 minutes [05:06] infinity, I'm getting the feeling from reading the README that localudebs/* are just tagged on to be installed during bootstrap, but won't be used for building the initramfs [05:08] psusi: Stuff in localudebs should end up in your mini.iso and be useable. [05:09] psusi: (The mini.iso is basically just isolinux, a kernel, and the d-i initrd, and not much else) === storm [i=joel@201.37.99.64] has joined #ubuntu-devel [05:09] infinity, yea... it will end up in the iso and it will be installed by anna during the setup bootstrap, but dmraid needs to be installed when the initrd is built, and I don't think it will be if placed in there === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [05:10] psusi: Oh, I see what you're saying. I also think you don't quite understand how d-i hardware detection works. [05:10] [05:10] psusi: At any rate, why would you care about dmraid running RIGHT AWAY? It just needs to run before partman. [05:11] psusi: Up to partman, we don't even care if we have a local disk at all. [05:11] the problem is most udebs are just installed by anna during the bootstrap... but dmraid has to be there when the initrd is built for it to be included... just having it installed after bootup won't work [05:11] because it doesn't "run" when installed anymore... I changed it to only set the initrd hooks, and it is expected to run in the initrd [05:12] of course... I suppose I could have the udeb built such that it runs dmraid when it is installed... [05:12] Fine for an installed system, probably a broken assumption for d-i. But you'd want to talk to Kamion about that further. [05:12] rather than setting the mkinitramfs hooks [05:13] hrm... yea.. that sounds like the right thing to do... have the normal deb just set the initramfs hooks, and have the udeb just run when installed [05:13] because anna can load some udebs that might be required to access the hardware after the initrd stuff has already run [05:13] psusi: Right. === psusi just felt his brain swell again [05:14] then again, that could be the beer [05:14] psusi: It will need to be run after (or as a part of) hw-detect. [05:15] ok... so I'm thinking it will need to be added to pkg-lists/access [05:16] and it's postinst should just run dmraid -ay, and the package itself should only include the executable... screw the docs and the mkinitramfs hook scripts [05:16] Right. [05:16] ok... cool... shouldn't be too hard to do.. [05:16] And, if you can futz with the build system enough to build the binary twice, the dmraid binary in the udeb should be compiled with -Os [05:17] Actually, if it's meant to end up in the iniramfs on installed systems, it might not be a bad idea to just compile the whole thing -Os anyway, and only do one pass. [05:17] Small initrds make people happy. [05:18] it's usually a good idea to use -Os anyhow as half the time it generates code that runs faster as well, due to fitting in the cpu cache better [05:18] ;) [05:19] but... that is another thing I've been wondering about initrds... early on in the live of that bug, Fabio mentioned klibc... [05:19] as far as I can see, klibc is NOT being used in ubunutu initrds [05:19] It is, but so is glibc. [05:19] In other words, it's bloated. [05:19] heh, yea, having both is stupid ;) [05:20] I'm rewriting bits of initramfs-tools to now select one or the other, based on your installed package set. [05:20] so I don't see any point in klibc then [05:20] especially since it seems to be a general practice to statically link to it [05:20] I say just put libc.so in there and everyone use it [05:20] (ie: we check to see if anything in your hooks requires glibc, if so, copy in the glibc-linked klibc-utils, and don't use klibc at all... If you don't need glibc, we use klibc exclusively, and get a smaller initrd) [05:20] That's the plan, anyway. [05:21] when it uses klibc, is it -shared? [05:21] and how much smaller is klibc.so anyhow? [05:21] The idea that klibc and EVERYTHING YOU MAY POSSIBLY WANT IN AN INITRD!! will be able to get along is a pipe dream we've given up on. [05:21] It's shared, and it's tiny. [05:21] like how much smaller than normal libc? [05:21] 5k? 25k? 1 MB? ;) [05:22] -rwxr-xr-x 1 root root 17020 2006-01-17 07:18 /lib/klibc-t2jM36h7OcxUNTDzncfER2p7kd4.so [05:22] -rwxr-xr-x 1 root root 1138716 2006-01-14 01:22 /lib/libc-2.3.6.so [05:22] WHOA!?!@# [05:22] how the hell they do that? [05:22] It's kinda featureless. :) [05:23] (or glibc is kinda feature-bloated, pick one) [05:23] haha, so it's not really usable then? :) [05:23] that's what I'm saying... is glibc just that bloated, or is klibc so stripped to the bone as to be hardly usable? [05:23] both [05:23] heh [05:23] It works well enough for the purpose of "tiny initrd, if you need it" Which was the point. [05:23] true... if all you need in there is busybox and a few kernel modules [05:24] so should I then have a klibc and non klibc deb and udeb? [05:26] I thought dmraid and klibc didn't get along right now anyway..? [05:26] don't get alone? the only thing about mixing them that I know of is it is rather stupid to add a smaller libc unless you can remove the larger [05:27] At any rate, for the forseeable future, we'll probabl have glibc in the initramfs, so I wouldn't bother. [05:27] ok [05:28] If you can test that it does build and work with klibc, that's cool too, I don't mind. :) [05:28] the upstream package has a --with-klibc configure option [05:28] also dietlibc I think [05:28] Having it doesn't mean it works. [05:29] it has some other options to you can use to strip out some less needed features for the udeb [05:30] btw.. is there a reason for using isolinux instead of grub on the cd? [05:33] Yes, it works better. [05:33] how so? [05:34] doesn't grub umm... totally kick ass? === b0rkb0rk [n=seth@d-ip-129-15-213-202.wireless.ou.edu] has joined #ubuntu-devel [05:34] Not for CD booting, apparently. === agabus [n=agabus@203-214-158-171.perm.iinet.net.au] has joined #ubuntu-devel [05:34] syslinux (and by extension, isolinux) is the king of booting removable media on braindead PC hardware. [05:34] psusi: AFAIK grub does not have a iso support module. [05:34] what difference does the medium make? [05:35] psusi: How the BIOS chooses to try to boot it makes a big difference. [05:35] psusi: And isolinux deals with a lot more broken BIOSes than grub does. [05:35] infinity: and syslinux looks prettier. :) [05:35] psusi: you need a standalone driver for the fs and the ability to understand what the BIOS is expecting of you [05:35] lifeless, so it can't boot el-torito native mode, put it and the kernel and initrd into a hd image and build it to use hd emulation boot [05:35] syslinux is king. [05:35] psusi: Back in "the day", grub was tried on the CD images, failed miserably in certain cases, and isolinux was used instead. === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === nnonix [n=nnonix@t42.bkjohnson.com] has joined #ubuntu-devel === seth|lappy [n=seth@d-ip-129-15-213-202.wireless.ou.edu] has joined #ubuntu-devel [05:37] those broken bioses will get you every time... when I was working on ReactOS I ran into a problem with the boot loader because it was written to use the bios disk number stored in the boot sector instead of the number the bios passes in CL [05:37] and in my case, the boot sector had 0 since mtools figures it's formatting a floppy [05:38] only the drive was 80 [05:38] turned out the reason the number exists in the first place and was being used from the boot sector is because a lot of bioses are broken and don't pass the boot disk number in CL [05:39] that was about as annoying as when I was working on the floppy driver and found out that no drives ever made actually support the detect disk command the standard called out for, which is why you have to tell the bios what kind of floppy you have, and the kenrel has to get that info from the bios and run with it... even if it is wrong [05:41] psusi: It's not "a lot", but yeah. grub only just started dealing with that [05:42] it pissed me off that because some idiot wrote a broken bios somewhere, the boot loader couldn't auto detect the boot medium on my machine, and had to rely on the hard coded value in the boot sector [05:43] I don't know how windows manages to boot from drives other than the primary master... or can it? [05:43] because fat boot sectors ALLWAYS say bios disk 80... unless it's on a floppy, then it's 0 [05:44] ntldr can chain to other partitions, but ntldr itself needs to be either on BIOS 0x80 (and on an active partition), or it needs to be chained from elsewhere (like grub) [05:45] and for grub to chain it, grub hooks int 13 to trick ntldr into thinking it's on disk 80 doesn't it? [05:45] Yes. [05:45] And that's all ntldr needs to be able to read far enough to read its config file (boot.ini), and from there it can start loading device drivers and boot "properly". [05:45] wait... does it hook int 13 and fake disk 80, or does it write 81 into the boot sector? [05:46] if you change the disk number in the boot sector to the correct value, it would probably work too [05:46] it's just that dos/windows allways formats hard disks with an 80 in there [05:46] If grub rewrites boot sectors when chaining, that's a bug, IMO. [05:46] (And I'm pretty sure it doesn't) [05:47] another stupid thing it does apparently is store the absolute lba sector number of the partition in the boot sector, and so the boot loader uses that rather than adding local offsets to the starting sector of the partition according to the MBR [05:47] so if you move the partition, you have to change the hidden sectors field in the boot sector for it to still boot [05:48] I seem to remember seeing some command in grub that has it rewrite the boot sector to switch active partitions [05:49] I'd have to look at my girlfriend's computer, where such an icky setup is in place. === jlje [n=agp@cpe-204-210-33-233.san.res.rr.com] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === psusi studies the hackery that is bootchart-udeb [05:53] Trying to find a suitable vmnet module for your running kernel. [05:53] The module bld-2.6.12-9-i686up-Ubuntu5.10 loads perfectly in the running kernel. [05:53] UNF! [05:54] WMware distributes pre-built modules for Ubuntu now? [05:54] Neat. [05:54] yay proprietary software [05:55] infinity: yeah, all we need is a .deb for player. [05:55] elmo: I suppose neon24 will need binary NEW as well, since those binaries were removed yesterday. :/ === jsgotangco [n=jerome@ubuntu/member/jsgotangco] has left #ubuntu-devel [] === jsgotangco [n=jerome@ubuntu/member/jsgotangco] has joined #ubuntu-devel === nnonix [n=nnonix@t42.bkjohnson.com] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel [06:25] Hah. VNC breaks because it can't find xfonts-base. [06:25] Because it got moved out of /usr/X11R6 [06:25] How in the heck to change this. [06:31] Ahh I see. It pulls it's config from /etc/X11/xorg.conf if it exists. === JaneW [n=JaneW@dsl-146-165-158.telkomadsl.co.za] has joined #ubuntu-devel [06:34] hi janew === pappan [n=garbage@bgepxyout-02.asiapac.hp.net] has joined #ubuntu-devel === mahangu [n=mahangu@unaffiliated/mahangu] has joined #ubuntu-devel [07:05] wow... I think I've udebified this package [07:06] hrm...probably should validate the depends... [07:15] hrm... when I try to build debian-installer it says it can't find libdebian-installer4 [07:17] well, it's late... I'll pick this up tomorrow... === stu1 [n=stub@ppp-58.10.73.144.revip2.asianet.co.th] has joined #ubuntu-devel [07:24] hi daniels === Lathiat [i=lathiat@ubuntu/member/lathiat] has joined #ubuntu-devel === raphink [n=raphink@c529d3992.cable.wanadoo.nl] has joined #ubuntu-devel === marilize [n=marilize@196.36.161.235] has joined #ubuntu-devel === irvin [n=irvin@ubuntu/member/irvin] has joined #ubuntu-devel === viviersf [n=cain@196.44.1.98] has joined #ubuntu-devel === Gman- [n=gman@nwkea-socks-1.sun.com] has joined #ubuntu-devel === highvoltage [n=Jono@196.36.161.235] has joined #ubuntu-devel === Den [n=Den@adsl-71-132-141-13.dsl.pltn13.pacbell.net] has joined #ubuntu-devel === freeflying_ [n=freeflyi@61.190.65.8] has joined #ubuntu-devel [08:07] Anyone know if today's dapper live iso built properly, no errors, etc? === JaneW [n=JaneW@dsl-146-130-233.telkomadsl.co.za] has joined #ubuntu-devel [08:08] Den: check the timestamp, if it hasn't built, it'll not be of today [08:08] Den: if it's workable / usable, no clue. [08:09] Den, there is an error log somewhere, maybe under people.ubuntu.com/~cjwatson [08:10] HiddenWolf: Thanks - the time stamp is new, but yesterday Mithrandir told me it didn't build properly, even though there was an updated time stamp. Thats why I asked. Thanks! [08:10]