[12:03] k. besides that it's looking pretty neat. just need to get me a kernel that does work with cdrecord, want the mouse to work this time [12:04] aha, that's why apt-ftparchive is sometimes so slow - changing the seeds changes the overrides which invalidates some of it's caching [12:07] mdz: I'm going to override Origin: to Ubuntu, and Bugs: to .. what? [12:07] elmo: mailto:some-address-get-it-from-justdave [12:07] ok [12:07] justdave: ping? === tvon [tvon@dsl093-119-225.blt1.dsl.speakeasy.net] has joined #ubuntu === tvon [tvon@dsl093-119-225.blt1.dsl.speakeasy.net] has joined #ubuntu === tvon [tvon@dsl093-119-225.blt1.dsl.speakeasy.net] has joined #ubuntu === tvon [tvon@dsl093-119-225.blt1.dsl.speakeasy.net] has joined #ubuntu [12:34] fabbione: yo (but too late, I know) [12:34] mdz: yo [12:35] Kamion: I think I was going to ask you if the current daily was any good, but I tried it and ran into the same problem jdub had (hdparm vs. SCSI/USB drive) [12:36] mdz: boot with cdrom-detect/cdrom_hdparm= on the kernel command line then [12:36] alextreme: when do you think the updated live CD can be ready? [12:37] i've got one ready now [12:37] just hasn't been tested much [12:37] mdz: p.s. having all man pages for a component or whatever in the one package would be SO MUCH CRACK [12:37] alextreme: let me at it; I'll give it some testing love [12:37] Kamion: :-) [12:37] live cd testing! [12:37] :) [12:37] YEAH! [12:37] LATHER UP! [12:37] okay, i'll do one final test and start uploading... which will take all night :) [12:38] TEARS [12:38] you should so get in the habit of building in the LAN and rsyncing down for testing [12:38] I started doing that right after the pain of uploading Sounder CD 1 from here [12:39] that was about when I commandeered little for CD building :) [12:40] yeah, i just need root (chroot) access for my autobuilder :) [12:40] you don't need root, we have dchroot [12:41] dude, he needs to run debootstrap [12:41] does warty have input methods outside of libgtk2.0-0 ? [12:41] oh, I thought he meant root to get in the chroot [12:41] that's so not cool then :P [12:41] Uclintu: yes [12:42] Uclintu: uim, xim, etc. [12:42] well, i wouldn't mind building in a chroot... :) [12:42] mdz: those appear to be in universe [12:42] hmm [12:42] alextreme: root in a chroot is root.. that's not a possiblity I'm afraid, not on little [12:42] mako: ping? [12:43] Uclintu: that's something we're looking at sorting out properly for Hoary [12:43] elmo: yeah, that's what i guessed :) [12:43] mako: I thought we added the input method stuff to desktop, but apparently not, do you remember what happened? [12:43] er, supported [12:43] elmo: the automated job is going to need root regardless [12:44] jdub: fair enough [12:44] mdz: gar - we're only doing this crack for one arch right? [12:44] elmo: for warty :-) [12:44] jdub: is that what happened? we decided to punt for warty? [12:45] -ECRACKWHORESANONYMOUS [12:45] dunno if it works already, but we're going to want install/live CDs for all arches :) [12:45] mdz: pretty much, being beyond freeze and knowing it'll be a lot of work [12:45] I think we need a bounty to de-root the stupid thing [12:45] jdub: I thought we had decided to add the couple of packages to supported so that people who already knew about them could get them easily [12:46] elmo: thats going to be an interesting bounty, and not one i'm willing to take up :) [12:46] elmo: fakechroot? [12:47] I thought elmo and lamont and I talked about this at the conference [12:47] this lappy is so quiet :) [12:47] kamion: don't think it was me [12:47] Package: fakechroot [12:47] Description: Gives a fake chroot environment [12:47] This package provides a library which overrides libc functions, so [12:47] it is possible to use root-specific tools without root priviliges. [12:47] . [12:47] In fake chroot you can install i.e. Debian boostrap, create developer's [12:47] environment and build packages inside chroot'ed system using standard [12:47] non-root user account. [12:48] oh, maybe we did, I think my concern was that the guy seemed to have forked debootstrap to make it work with it [12:48] elmo: yes, only one arch, there's not going to be time to get amd64 and powerpc working AFAIK [12:48] jdub: we decided WAY back only to do i386 for warty [12:49] Kamion: yeah, but in-future-ideally [12:49] sure, yeah [12:49] I doubt amd64 is hard, powerpc might be interesting [12:49] (but, hey, largely the sort of thing I do ...) [12:49] UML is probably the way to do it [12:49] why not fakechroot? it exists, it's simple, it (allegedly) works [12:50] then we can (allegedly) use it [12:50] i'll give it a spin, see if it works here [12:50] does fakechroot let you mount()? [12:50] doesn't debootstrap mount /proc etc.? [12:51] ah, well, yes, fakechroot does talk about doing 'debootstrap --fakechroot', hmm [12:51] UML opens the possibility of actually booting and testing the thing [12:51] mdz: he was talking about ignoring mount/umount and /proc issues back in apr 2003 [12:52] setup_proc () { [12:52] - on_exit "umount $TARGET/proc" [12:52] - umount $TARGET/proc 2>/dev/null || true [12:52] - in_target mount -t proc proc /proc [12:52] +# on_exit "umount $TARGET/proc" [12:52] +# umount $TARGET/proc 2>/dev/null || true [12:52] +# in_target mount -t proc proc /proc [12:52] + for i in cmdline cpuinfo devices filesystems loadavg meminfo misc modules pa [12:52] rtitions stat swaps uptime version; do [12:52] + cat /proc/$i > $TARGET/proc/$i [12:52] + done [12:52] + mkdir $TARGET/proc/1 [12:52] } [12:52] hm, ok [12:52] that's a bit gross === bdale [~bdale@18-pool1.ras11.coden-le3.alerondial.net] has joined #ubuntu [12:53] mdz: and it closes the possibility of having a single build strategy on multiple architectures [12:53] eewwww [12:53] mdz: unless somebody actually finishes porting it to powerpc [12:53] Kamion: it gives us 2 out of 3 architectures [12:54] 2 out of (3+however-many-derived-distributions-want-to-do) [12:54] s/distributions/architectures/ [12:54] I did mean "derived distributions", not "derived architectures" :-) [12:55] then I don't understand how you arrived at that expression [12:55] however many [architectures] derived distributions want to do [12:56] http://lists.debian.org/debian-devel/2003/04/msg00747.html [12:56] ah [12:57] elmo: do you need #270906 for the bug stuff? [12:58] I wonder how many bits of what debootstrap does actually require real root [12:59] I could imagine having those allowed by extremely limited sudo or a userv service or something [12:59] probably not much beyond chroot/setuid/permissions type stuff [12:59] ability to untar archive containing device nodes, hm, that's probably about it ... [01:00] the last is a bit hard to limit though, although the tarball in question is in /usr/lib/debootstrap/ so you could assume it to be safe [01:01] chroot is obviously too much though [01:01] mdz: no, I can script around it trivially enough, it'd just be nice to some day not to have to [01:02] (thus wishlist) [01:02] shouldn't be too hard to implement [01:02] someday [01:02] UML :) [01:03] how about if the debootstrap-running script were separate and trusted? I thought it was going to run on the buildd anyway [01:03] kamion: yes, it'll have to be [01:03] well, it's wrong to have it on little for reasons other than security anyway [01:03] it's certainly going on the buildd rather than say little, but that doesn't mean I'm happy with it having root [01:04] sure :-/ [01:04] err.. [01:04] are these Xserves UP or are we running a UP kernel? [01:04] IIRC they're SMP but we're running a UP kernel [01:04] because getting the UP kernel up was hard enough without trying to deal with SMP too [01:04] cool [01:04] so out of our 3 architectures, only 1 actually works SMP [01:05] I think thom said he'd try to take one down (ross?) and see if the SMP kernel Just Worked [01:05] but dunno if that happened [01:06] mdz / elmo: pong [01:06] speaking of kernels working... [01:06] want any other info for #496? [01:06] that an address for bug reports you're looking for? [01:06] justdave: yeah [01:06] Uclintu: that bug's way over my head, I'm afraid, dunno about anyone else [01:06] unfortunately it's the sort of thing where sitting at the console and screwing around is about my only chance ... [01:07] we don't have one yet. I don't think we can do inbound email on macquarie, so we probably need to create one somewhere. [01:07] justdave: we need it very soon [01:07] Kamion: let me know when you're dropping by [01:07] Uclintu: might be worth hacking the initrd script to run with sh -x, if that's feasible === Gman [~Glynn@amfea-proxy-2.sun.com] has joined #ubuntu [01:07] then we could at least see WTF's going on [01:07] got a way to forward the mail to macquarie from whatever machine gets it, or want macquarie polling for it? [01:07] which initrd script? [01:07] /sbin/init in the initrd [01:07] justdave: ideally macquarie would poll [01:08] I'm assuming it's a shell script [01:08] Kamion: yeah; #!/bin2/sh [01:08] I can set something up on ubuntu.com, if needs be, just tell me what === Uclintu curses cramfs. [01:09] probably want something like submit@bugzilla.no-name-yet.com [01:10] does the second arg you pass to pivot_root have to exist before you call it? [01:10] pivot_root . initrd === Gman_ [~Glynn@amfea-proxy-2.sun.com] has joined #ubuntu [01:10] Uclintu: does your root filesystem have an /initrd directory? [01:11] Uclintu: yup, empty [01:11] give me a pop3 mailbox I can get at from macquarie or something [01:11] Uclintu: hm, in that case it suggests to me that it's mounting the wrong root filesystem somehow [01:11] Uclintu: oh, what's the $Id$ at the top of /sbin/init in your initrd? [01:12] # $Id: init,v 1.56 2004/05/18 10:48:15 herbert Exp $ [01:12] yes, that's the one [01:12] justdave: okay, adding that to my todo list [01:12] I think I'd check the value of root= [01:13] in yaboot.conf? [01:13] /dev/hda7, I see [01:13] yup [01:14] sh -x should tell you what it's actually trying to mount, anyway [01:14] okay, i'll replace the initrd and reboot [01:15] that Bugs: header is in the .deb file for packages? [01:17] justdave: it's in the Packages index file [01:17] justdave: it may or may not be in the .deb; in this case I think it typically won't be, it'll be overridden in the archive maintenance scripts [01:17] ok. [01:18] probably won't help me anyway because they mostly point at bugs.debian.org probably. :) [01:18] submit@bugzilla.ubuntu.com rather than bugzilla.no-name-yet.com perhaps? [01:18] I was looking for an easy way to find out who the upstream is for a given package [01:18] dunno what the rules for that are right now, but we need something we can wire into packages [01:18] /usr/share/doc//copyright [01:18] there's no machine-parseable way to get that information [01:19] do we have ubuntu.com now? I thought we weren't supposed to use it yet [01:19] we're not supposed to use it for web content ... elmo, is mail there ok? [01:20] kamion: I think so, yeah [01:21] [if it's not, not hard to change the overrides..] [01:21] yeah, i was gonna try the SMP kernels when we didn't need all the buildds for a while :-) [01:21] TMBG were frickin' excellent live === edd notes it's about time he saw thom again [01:22] GAR, need new chair! [01:23] edd: come to london, we can do malaysian again :-) [01:23] thom: i should do. i was just down there at the weekend. [01:24] mdz: k, it's still a little rough (duh) and uses 2.6.7. had to revert to the old kernel, 2.6.8.1 was giving overlaying problems [01:25] but it's uploading, so i don't want to think about it anymore :) [01:25] yaaaay === Uclintu [~clint@user-12hdtek.cable.mindspring.com] has joined #ubuntu [01:25] yup [01:25] isn't mounting real root [01:26] time to watch some mindless furutama and purge the livecd wackyness from my brain [01:26] gnight [01:28] Why am I plagued with the idea that hda may have transformed into hdc? === Keybuk [scott@descent.netsplit.com] has joined #ubuntu [01:29] mdz: input method stuff shoudldn't go in desktop yet [01:29] mdz: there are just too many of them and each one writes interfaces to all the others [01:30] mdz: supported, yes [01:30] mdz: scim, uim, and m17n are the ones i like [01:32] mdz: i'd *really* prefer to have users who are inputting text though tell us what they're using though.. i just don't have a sense for what is actually being used or even useful and it's hard for me to test [01:33] there's some freedesktop unification work going on atm [01:33] not sure if that'll be hoaryable, but it would be nice [01:33] I use uim these days, but I'm not totally pleased with it. [01:34] jdub: well scim and uim are both in freedesktop now [01:35] jdub: i THINL everything is going toward scim at the top [01:35] Urgh. Input managers. [01:36] It would be nice if X had some method for inserting arbitrary UTF-8 characters [01:36] It'd remove a lot of the pain [01:37] there's plenty of pain left [01:37] it's way better than it used to be [01:37] oh sure, absolutely [01:37] back when you had to use particular locales for each IM [01:37] mako! [01:37] Oskuro: DUDE! [01:38] mako: so where's the OLIVE STUFF! [01:38] doesn't GTK+ have all of its own IM stuff as well? [01:38] Oskuro: http://mako.yukidoke.org/copyrighteous/ === ddaa [~david@nemesis.xlii.org] has left #ubuntu ["Client] [01:38] Oskuro: i just put it up like 2 minutes ago [01:38] mako: oh so it's there? [01:38] oh, lol [01:39] I looked at planet [01:39] Kamion: I wonder if there's some kind of ubuntu d-i big tarball where I can look for the ca.po's. [01:41] "The went into more detail" [01:41] typo there [01:43] gotta go to bed. === Oskuro pokes Kamion === Uclintu [~clint@user-12hdtek.cable.mindspring.com] has joined #ubuntu [01:47] well, that would explain it [01:49] Keybuk: gtk has some stuff im stuff, but not very advanced stuff [01:52] Uclintu: hda/hdc have been known to swap due to ATA-100, but I thought that problem was gone [01:52] Oskuro: no, not yet [01:52] Kamion: the real problem is that it's only finding my second IDE controller [01:52] Oskuro: you could probably walk /dists/warty/main/debian-installer/binary-*/Packages.gz reasonably automatically ... [01:52] Uclintu: oho [01:52] i'm going to force a "modprobe -k cmd64x" in loadmodules and see if that helps any [01:52] Uclintu: is there something unusual about the first one? [01:53] it's a cmd646 [01:53] wow, elderly [01:53] does the installer work on that? [01:53] daniels: ping [01:53] I don't remember if I've ever tried an install. [01:53] you don't actually have to overwrite anything on the disk, just boot the CD, run through the UI up to the start of partitioning, and see if it displays all your disks [01:54] jdub: huh, preview release is Wednesday, not Monday? [01:54] i've only got sounder 4 burned; that relevant? [01:54] Kamion: yes, always has been [01:54] mdz: somebody needs to update http://wiki.no-name-yet.com/WartyWarthog not to lie, then [01:54] Uclintu: no, it really isn't. please do try a sounder 8 though [01:54] Uclintu: well, that's only 2.6.7, but it would be a useful data point [01:55] Kamion: that indicates the week in which it is scheduled [01:55] it is admittedly misleading [01:55] okay, first my loadmodules hack [01:55] mdz: it's so not helpful :) [01:55] Kamion: feel free :-) [01:55] and gold release is Oct 15th? [01:55] correct [01:56] updated [01:57] "anal-retentive, high-caution period" [01:58] um, the date has always been correct though [01:58] I guess this week has been relatively cautious compared to last week [01:58] what's the official release date? [01:58] debianist: preview on the 15th [01:58] jdub: the WartyWarthog page did not have those dates anywhere on it that I could see [01:58] of october? [01:59] september [01:59] jdub: only the date of the preceding Monday [01:59] final in october [01:59] jdub : and security review is due by the 15th or the final in october? [01:59] Kamion: yeah, that's enough, thanks [01:59] night all [02:00] debianist: it's just a running process === Uclintu [~clint@user-12hdtek.cable.mindspring.com] has joined #ubuntu [02:01] night [02:02] so, forcing "modprobe -k cmd64x" into loadmodules, then booting with "Linux root=/dev/hdc7" got me here, after several screenfuls of ieee1394 spew and a lack of the ethernet driver being loaded [02:03] should I record this in the same bug, a different bug, or unspecified action? [02:03] Kamion: (also informed by the date of the gnome release, which, should it slip, is the date of our release) [02:03] debianist: the initial review needs to be done by preview, yes [02:03] jdub: how is that looking, by the way? will gnome make it? [02:04] yes [02:04] Uclintu: same bug I think [02:04] unless something completely bizarre happens, it's looking fine [02:04] jdub: wanted to run a couple of menu ideas by you [02:05] mdz : noted. guess not much sleep for the coming week :) [02:05] jdub: what do you think about removing ggv from the menu? it's not particularly useful to run by itself, even if the user does have postscript files to work with [02:06] I don't think any of us will be getting much sleep between now and preview :P [02:06] assuming there is a way to hide it from the menu and still let the mime association stuff work [02:06] I warned my gf today that I'd probably be going into hiding [02:07] mdz: there are a number of file handling apps like that, i don't think we should remove them from the menu at this stage of the release (!!!) because there's no other way for users to find/run it beyond having files already [02:07] jdub: that's just the point: running ggv is damn near useless [02:08] mdz: the menu is currently like /Applications; it's not a convenience, it's everything [02:08] mdz: i don't think it's appropriate to make what are essentially UI policy changes like that at this end of the release process [02:09] I had a feeling you'd say something like that [02:09] mdz: that said, it is definitely possible to do it without affecting the mime system (you just pull the categories line out) [02:09] well come on [02:09] we're already fucking around with all kinds of other things [02:10] i don't like making UI changes without useful design/thought/testing [02:10] we're hardly finding / dealing with the side-effects of the other things we're changing [02:11] and totally abusing the release process as we go [02:14] release process? wassat? [02:14] jdub: we're not talking about software engineering here; this is totally harmless stability-wise, and I asked for your opinion. there is no need to be hostile about it [02:15] it's hardly any different than deciding to kick something out of desktop into supported, which we just did, with your emphatic agreement, a couple of days ago [02:15] and that's the next best thing === Uclintu [~clint@user-12hdtek.cable.mindspring.com] has joined #ubuntu [02:17] that did not go well [02:17] Uclintu: cd test? [02:18] yeah, it loaded ide-scsi before cmd64x, then spent an inordinate of time playing with sd_mod before I lost all patience [02:20] ew, ide-scsi [02:20] whose fault is that? [02:22] mdz: not hostile [02:22] discover (0.9.10) unstable; urgency=low [02:22] * Hard-coded a special case: add ide-scsi as a SCSI driver if IDE CD-ROM [02:22] drives are present. This way, all CD-ROMS will be treated as SCSI, [02:22] and burners will work out of the box. [02:23] I'm guessing [02:23] Kamion: does d-i rely on that? [02:23] I'd be surprised [02:23] although who knows :-) [02:23] if not, we should revert that [02:23] Kamion, but last time I checked, sg.o wasn't loaded, so it didn't work ANYWAY [02:23] mdz: stability is also a UI issue, not just a code issue [02:24] Hrdwr_BoB: quite, but that's a separate bug which I've already replied to [02:24] Hrdwr_BoB: ide-scsi is neither necessary, nor a good idea, for writing CDs anyway [02:24] yeah, agreed [02:24] i had to load sg manually too [02:24] Uclintu: could you file a bug about the ide-scsi thing? I'll fix it tomorrow [02:25] jdub: hence #1037 === jamesh [~james@203-59-250-175.dyn.iinet.net.au] has joined #ubuntu [02:26] anyway, I so have to go; night all [02:26] Kamion: #1128 [02:26] so, replacing dsdt crack - is that fixable in a general way? [02:26] Uclintu: thanks [02:26] jdub: dsdt's are machine specific [02:27] I've just given up on CD writing [02:27] I just write DVDs with growisofs [02:27] which does not require ide-scsi madness *or* scsi generic madness [02:27] AIUI you need someone that knows the acpi assembly to review the dsdst, fix it, and then it can get loaded into your kernel as a fix. [02:28] I don't know if there is a database-looked step (machine x, oh it needs *that* override) or not. [02:28] I have two laptops here that badly need dsdst loving though :| [02:33] is it something that a bios upgrade could fix? [02:42] acpi.sf.net has a collection of dsdts for specific machines + bioses [02:42] there's one there for my machine/bios combination [02:42] but it requires patching the kernel [02:42] which doesn't seem particularly generalisable [02:43] (the way it's patched, the the requirement for patching) === jdub wonders how his lappy handles mediabase hotplugging... === StoneTable [~stone@c-67-167-111-186.client.comcast.net] has joined #ubuntu [02:46] I thought the load custom dsdt stuff was quite generic? [02:46] you just compile it into your kernel [02:46] by excluding the existing foo [02:46] (it seems) === Gman_ [~Glynn@amfea-proxy-2.sun.com] has joined #ubuntu [03:04] mdz: pingmdz pong [03:04] um, yeah [03:05] that === haggai_ [~halls@pD9EA624A.dip.t-dialin.net] has joined #ubuntu [03:49] daniels: going to take care of #1056 today? [03:51] jdub: could you please subscribe lsd@linuxgamers.net to sounder? he tried to subscribe a while ago but never heard anything [03:51] mdz: yes, hopefully [03:51] mdz: i have a dist-upgrade configuring shit now, and maybe tetex will work after this one === bdale [~bdale@216-43-77-134.dsl.mcleodusa.net] has joined #ubuntu === cef_work [~cef@fw-01.amc.com.au] has joined #ubuntu [04:44] hrm, I just thought of something after noticing the xchat stuff. we're going to end up with people in here asking general questions. We're probably going to need another channel for devel stuff, and also we might want an infobot (like dpkg & apt on #debian) for the common questions and answers [04:55] heh [04:56] i could never learn dvorak unless i actually had a dvorak keybaord [04:56] but they dont make a microsoft natural elite in dvorak :) [04:56] i keep on wanting to get a frogpad [04:56] tho [05:04] 44 === Riff [~davyd@cook.theducks.org] has joined #ubuntu [05:14] now i remember why i hate dvorak [05:14] i have to fuck with vi to get my vi keys back :) === stub [~zen@dialup-30.83.220.203.acc06-dryb-mel.comindico.com.au] has joined #ubuntu === kagou [~kagou@AMontpellier-251-1-30-53.w83-113.abo.wanadoo.fr] has joined #ubuntu [05:22] hi [05:23] man, I'm not geeky enough [05:23] I thought you meant the composer [05:24] jdub, have you seen my request ? [05:24] kagou: hrm, no? [05:24] i had written a too big mail to the list, waiting for your approbation [05:25] aha [05:25] :p [05:25] ""Release8 reportinstallation" is the subject [05:26] done [05:26] thanks jdub :) === jdub is in luxury with fast new laptop [05:28] ^^ [05:30] jdub: If only you could have made the toilet seat faster, you'd have the best of both worlds... a hoverseat! ;) [05:30] jdub: did you patch your DSDT ? [05:31] Riff: stopped given business and inability to compile iasl [05:31] might ask steven if he already did his [05:31] utuntu :: now with Dell support! [05:32] i don't think the dsdt hacking stuff is very generalisable though [05:33] it all gets merged upstream eventually [05:34] the linux ACPI code is full of workarounds for specific machines and fuck knows what else [05:34] this is apparently why the BSD guys refuse to support ACPI [05:34] Riff: what will they do when there are no more machines their OS will run on? [05:35] jamesh: I have no idea [05:35] maybe they blissfully believe IBM will have APM forever [05:35] I mean, this laptop doesn't have APM [05:36] if you have working APM, is there any benefit in trying to get ACPI working? [05:36] jamesh: not unless you're trying to do fucked up things I guess [05:36] ACPI is infinitely configurable [05:37] it's a much better design from the OS<->Hardware layer [05:37] if only people didn't make mistakes in implementing it [05:37] APM is all hardware, so your OS has to do a lot of work to cope with it === sabdfl [~mark@wblv-228-55.telkomadsl.co.za] has joined #ubuntu [05:37] there are also lots of magic weird BIOS routines to get APM to do things [05:37] ACPI doesn't have that [05:38] ACPI tells you a button was pressed, you do all the work yourself to prepare for suspend state [05:38] then you tell ACPI to suspend [05:38] "The world's biggest ice-cream maker plans to genetically modify ice blocks in an attempt slow down the melt rate." [05:38] what exactly are they genetically modifing? [05:39] because that one sounds a little scary === Riff is all for GM food usually [05:39] http://www.smh.com.au/articles/2004/09/09/1094530771695.html [05:43] I'm sure everyone would support GM if it could keep their beer colder longer [05:43]