[12:02] jdub: happy with a dbus&hal update at this point ? [12:03] seb128: if they go in soon, and mdz is happy with it, then it's a necessary evil :) [12:03] ok, I'll look on this tomorrow so [12:04] cool [12:14] what's a good name for the udeb spat out by shadow that asks questions about the initial user account to create? [12:14] it's kind of a clone of passwd.postinst but doesn't have most of the other code in passwd [12:15] adduser ? [12:15] initial-username.udeb? [12:15] hmm, no. too confusing [12:15] and deb/udeb names share a namespace anyway [12:15] initial-user.udeb since it does more than just a username [12:16] if it ever went into Debian it would in principle ask all the questions from passwd.config, which has stuff like "do you want md5 passwords?" [12:16] passwd-config.udeb? [12:16] WFM [12:17] CONSENSUS [12:17] woohoo [12:18] jdub: would you create ubuntu-security-announce? [12:19] ok [12:19] Kamion: *g* [12:22] hm, I think I shall delete the root password question entirely from passwd-config.udeb; even if you wanted to set a root password, storing it in the debconf database across reboot would be evil! [12:24] mdz: for tbird i'll just upload a port of new debian with our additions for amd64 tomorrow; i don't think a backport is beneficial right now. waddya think? [12:27] when ubuntu service pack 2 will be released? [12:27] lol [12:27] ups, sorry, that's devel channel [12:29] of course, the user password ends up in the debconf database over reboot ... [12:29] mdz: you have a problem with that? it'll be deleted as soon as the user's added [12:31] Kamion: config.dat or password.dat? [12:32] mdz: there's no such distinction in cdebconf [12:32] sorry, yeah, the *c*debconf database over reboot [12:32] Kamion: :'-( <-- baby jesus [12:33] yarrrrrrr [12:33] Kamion: is the cdebconf database world-readable? [12:33] if so, can we make it not so? [12:33] /var/log/debian-installer/cdebconf/questions.dat is world-readable post-install [12:33] it would be a pain in the arse to make it not so, since it's useful for bug reports [12:33] is there any reason why it needs to be? [12:33] can't you save password md5'ed ? [12:33] there is other stuff in /var/log/debian-installer which is not world-readable [12:33] like syslog [12:33] but we could make it not world-readable if we were willing to suffer the support load [12:34] trukulo: not without extending adduser to support that === ddaa [~david@nemesis.xlii.org] has joined #ubuntu-devel [12:34] mdz: ok [12:34] actually, passwd.config doesn't use adduser for that part [12:34] it does adduser --disabled-password, then its own evil chpasswd hack === ddaa points the Monty Python finger and scream "splitters!" [12:35] mdz: would you be happy for a *hash* of the password to be in a world-readable file? I'm uneasy about it myself [12:36] I don't think the unencrypted password should be in any file, world-readable or not [12:36] Kamion: not really [12:36] but I'd be even happier with a hash in a non-world-readable file than with a password in a non-world-readable file [12:36] I think the latter's right out [12:36] we might as well store /etc/shadow unencrypted at that point [12:37] that's what we'd be doing with the former [12:37] what about saving password not in database, but in a different file and then remove it later? [12:37] no, I mean the passwords in /etc/shadow unencrypted :) [12:37] trukulo: too much work for the time period [12:37] Kamion: me too :-) [12:37] Kamion: right [12:38] I have no qualms about making questions.dat root-only if we're going to put authentication data in there [12:38] trukulo: the only way this is even feasible in the TWO DAYS I have to implement this is to use debconf :) [12:38] we can solve the support issue if needed [12:38] all right [12:38] I'm still a little queasy about putting the password in there [12:38] then I have to chroot to run perl === mako [mako@micha.hampshire.edu] has joined #ubuntu-devel [12:38] only two days? then non-readable, as you said before [12:38] non-readable hash is equivalent to /etc/shadow, clearly ok [12:39] but if you accomplished that, then people don't need to restart after installation? going into X automatically? [12:39] they still need to reboot after the first stage [12:39] I don't think that's negotiable; you need to be sure that the boot loader works [12:40] but no passwords at first boot after install, that's very usable [12:40] Kamion: i agree, i prefer forcing to reboot after configuration [12:40] [personally I don't really see that it's much different from the current case, but the guy who writes the paycheque wins, as lamont sometimes says ...] [12:41] :) [12:41] little diferences like these, makes a good look for users [12:41] I don't even see the usability win myself :) [12:42] nor do i, but remember that most people thins graphical installers are better than ncurses === Kamion ponders setting a 'hash' flag on passwd/user-password once the password is hashed [12:43] to distinguish [12:43] Joey Hess has observed in the past that I'm one of the few people who uses custom debconf flags in maintainer scripts ... [12:43] (with a reaction like "wow, you actually do that?", IIRC) [12:44] i'm not a programmer, i'm only sysadmin :) so i can't understand what you talking 'bout very well [12:44] just thinking out loud [12:45] that's good technique [12:45] when we talk bout we thinking, we're thinking better [12:45] k, i'll shut up, don't want to make noise here [12:45] jdub: ubuntu-security-announce ready? I'd like to follow up with a hyperlink to subscribe [12:46] sec [12:46] sorry, didn't think it was immediate request === jdub is eating breakfast;) === trukulo thinks bout going to bed 0:40 here [12:47] # Use perl rather than echo, to avoid the password [12:47] # showing in the process table. (However, this is normally [12:47] # only called when first booting the system, when root has no [12:47] # password at all, so that should be an unnecessary precaution). [12:47] SETPASSWD_PW="$2" [12:47] export SETPASSWD_PW [12:47] surely environment variables show up in the process table too? [12:48] hm, perhaps only for processes you own [12:49] mdz: explicit reply-to ubuntu-devel? [12:49] jdub: ubuntu-users, please [12:50] Kamion: under Linux, yes, only for processes you own [12:50] a poor feature to rely on for security [12:52] yeah, fortunately not in fact relied on [12:53] kind of hard to deal with sensitive data securely in shell [12:54] doogie probably wrote some oo-code for that [12:55] doogie and secure in the same sentence? :) === ddaa [~david@nemesis.xlii.org] has left #ubuntu-devel ["Client] [01:05] hm, no registry of prebaseconfig item numbers [01:29] elmo: ping? === __randy__ [~randy@sclab-25-433.sclab.clarkson.edu] has joined #ubuntu-devel [01:37] daniels: ping? [01:41] Kamion: I didn't realize that had become "my" statement... [01:43] lamont: build errors coming in :| [01:43] oh, needs NEW flushing [01:43] nm [01:47] fabbione, thom: most recent firefox and xfree86 both have no problems for me on all architectures, thanks! [01:47] jdub??? [01:47] lamont: did some new stuff for universe+hoary testing late last night [01:47] fucked up a bit [01:48] i am a bad person [01:51] elmo: ping? [01:53] gah. so when one changes Makefile.am, just need automake, or autoconf as well>> [01:54] automake [01:56] lamont: you're who made it familiar to me :) [02:02] Kamion: ah, ok [02:02] also know as the Golden Rule: he with the gold makes the rules. [02:03] BLING! [02:03] bling-bling [02:09] Is it frightening that I know the PCI vendor id for Broadcom by heart? [02:22] elmo: ping [02:30] Kamion: there was a time when I could ID many networking cards from the first 3 octets... === Kamion hands lamont pci.lst; memorize that, I'll test you tomorrow [02:31] ENOTIME [02:31] oh, ok, you can have until Monday then [02:32] would certainly make my life easier if I knew the whole lot by heart :) [02:33] Kamion: I've been known to really annoy people by assembling parisc instructions in my head.. [02:33] mdz/jdub: ? [02:33] repeat after me: 8 7 4-1 6 5 19-9 0 [02:33] I think that's the order, anyway... [02:34] s/annoy/freak out/ === lamont has to run the kids to town [02:34] elmo: yeah [02:34] jdub: is this NEW stuff for universe or main? [02:34] the bad part is when you patch about 6 instructions in one fell swoop, in hex. [02:34] back in a couple of hours. [02:34] elmo: universe [02:34] elmo: can you kill polypaudio in new? [02:34] elmo: i'll fix that up [02:35] libao-polyp too? [02:35] yeah, ok [02:36] remember, universe stuff needs to be universe/$SECTION... I can override tho, so it's not critical [02:36] oh [02:36] libao-polyp (0.3-1) universe/sound; urgency=low ? [02:37] or in control? [02:37] yeah, but you should change debian/control - not the .changes [02:37] so Section: universe/sound -> for both binaries and source? [02:37] lamont: you're insane, dude :) === jdub chuckles at elmo's reject mail [02:39] jdub: yeah [02:39] jdub: share [02:39] "Rejected at request of bong sipping maintainer [02:39] " [02:39] :-) [02:39] we should patch our vim to understand our wacky debian/control + changelog foo [02:40] ;) [02:40] oh yes [02:41] jdub: so am I rejecting them all? [02:42] just those ones [02:42] ready to upload? [02:42] yeah [02:42] rocking, thanks [02:44] did I ask whether you guys are keeping all your main debian/ dirs in a repository? [02:44] if not, do you? =) [02:45] we will [02:45] there's some interesting stuff on the way in that regard :) [02:45] sweet [02:49] (new) polypaudio_0.5-1ubuntu1.dsc optional sound [02:49] is that intentional? [02:50] not universe/sound? no, i did that one before you mentioned it [02:50] ok, no prob, I'll just override [02:50] I assume gamin/howl aren't going to main at this stage either? [02:51] no [02:51] thanks [02:51] i'll do that in future :) [02:52] elmo: upload access for herbert? [02:52] thom: ick, does libapache2-mod-php4 really only work with the prefork mpm? [02:52] well, meh, does anyone have any suggestions on an upload method that allows you to reliably tell when a file's completely uploaded? [02:53] keep rsyncing until it doesn't do anything? :) [02:54] elmo: run a twisted ftp server, and fire off events when the files are fully uploaded :-) [02:54] sheesh, new Windows XP Home is 163? [02:55] not buying that just for testing purposes [02:55] maybe one of my friends can lend me a copy or something [02:57] expense it ;-) [02:57] "required for the express purpose of bringing freedom to the entire world" [02:58] tempting [02:58] "(also, doom3 looks very interesting)" [02:59] actually, very tempting; then I can continue Not Owning A Copy Of Windows [02:59] ('cos it won't really be mine) [02:59] i didn't get that badge ;) [03:00] are windows XP licenses transferrable? [03:00] if so, you can have the one that came with my laptop [03:00] (they're definitely not with the hardware) [03:00] (dunno about otherwise) [03:01] oh [03:01] lwn [03:01] not legally, but they work === jdub goes to check out ubuntu article [03:01] boh [03:01] not front page [03:01] " Here is a little quiz. Which Linux distribution's mailing list recorded over 1,000 posts during the first week of its existence? Which project succeeded in attracting some of the best-known and most prominent open source developers to work on it? And why do their email addresses invariably end with @canonical.com?" [03:01] so what's all that product activation about then? [03:01] haw haw [03:01] I thought it was supposed to stop you transferring copies around? [03:02] kamion: it definitely doesn't, there's a key that's assigned "to the NHS" that must be used on thousands of computers [03:03] -rw-r--r-- 1 root root 2266975 Sep 22 15:42 changelogs [03:03] that's the apt-listchanges output for a woody->warty upgrade on a non-trivial server [03:03] neat [03:03] that's a hell of a lot of changelogs [03:04] oh, so Dave Miller really is *the* Dave Miller. Should have figured. [03:04] uh? [03:04] no, our Dave Miller is not the kernel guy [03:05] then lwn is wrong [03:05] neat [03:05] heh [03:05] I should do that subscription via debian thing sometime [03:05] everbody's been named except elmo === jdub sends correction [03:06] plenty of Debian people other than elmo not mentioned either [03:06] azeem: everyone? :) [03:06] Hi, [03:06] Dave Miller on our team is not Dave "dude" Miller, it is Dave "bugzilla [03:06] maintainer" Miller. :-) [03:06] Thanks, [03:06] - Jeff [03:06] [03:06] ;-) [03:06] heh [03:06] namechecking me as "Debian QA" is a bit out of date too, but I won't argue :) [03:06] jdub: ITYM, "K, Thanks, Bye. Love, Jeff" [03:07] haha [03:07] the other dave miller says dude *way* too much [03:07] too many Dave Millers out there :) [03:08] as far as I know, 'davem' still works for Red Hat [03:08] yeah, he does [03:12] oh, lwn is out? I guess they go by UTC [03:21] http://unconcerned.org/index_unbutu.html [03:21] hahahahaha [03:21] (for non-gnome dudes, see goneme.org) [03:29] :-) [03:36] new d-i uploaded with slightly branded installation manual, still lots of work to do there [03:36] shout if it fails to build for some reason; I don't think it will, but ... === Kamion is off to bed === jamesh [~james@203-59-87-199.dyn.iinet.net.au] has joined #ubuntu-devel [03:50] Kamion: night [04:01] my god, the "submit a bug" page is half a mb big? [04:07] most of that is probably the component list. [04:14] mdz: pong [04:14] so [04:14] sbuild or pbuilder [04:14] ? [04:14] elmo: dude, that's a feature [04:14] elmo: dan jacobson can never report a bug [04:14] jdub: none of the above [04:15] mdz: what do you recommend? [04:15] mdz: LOL [04:15] jdub: hypothetical UML-based tool I've only half-written [04:15] jdub: for what? === jdub spanks mdz [04:15] elmo: for building and checking things locally [04:15] pbuilder is scary crack and sbuild isn't particularly well suited for non-buildd use [04:15] sbuild is machosistic, unless the package is vastly improved [04:16] pbuilder's always struct me as particularly bad crack, but I don't have much experience to back that up [04:16] i basically want something so i can build shit against buildd-cletn unstable or warty whenever [04:16] elmo: of course it's vastly improved, I'm co-maintainer now [04:16] didn't do anything yet, though [04:17] heh [04:17] personally, I just use a chroot by hand, but I'm 0ld sk00l that way [04:19] how do you uninstall the Build-Deps afterwards? [04:20] *cough*, I cut'n'waste the "the following packages will be installed" output from apt-get to a file, and then apt-get --purge remove $(cat to_remove) afterwards [04:20] haha [04:20] that's not old skool [04:20] that's no skool [04:24] hey, I wrote the precursor to 'sbuild' in a hideously eccentric little shell called 'es' (-> 'esbuild') - I get to be a freak about how I build packages in a chroot [04:30] jdub: module.h:26:18: ltdl.h: No such file or directory [04:30] missing b-d on libltd3-dev or whatever it is, I guess [04:31] yeah, now they're out of new i can upload the fixed one :) [04:32] I just throw the chroot away afterward === jdub feels morbid, migrates to pillowful working area [04:40] hmm, no hoary/supported seed yet? [04:41] hoary should be sharing warty's seeds for now [04:42] yeah, I mean for proposals [04:42] as I assume it's too late to propose things for warty :) [04:43] 8) [04:43] mdz/jdub: btw, do you have a timescale when you want hoary, if it's not ASAP? [04:44] elmo: to be honest, I don't see the point until we have hct [04:44] elmo: I don't think it's too late to propose stuff for supported if it's trivial [04:45] mm.. Mark indicated to me that hct was a way away, e.g. he said we'd be accepting community uploads before hct was ready, and I think that included hoary [04:45] I spoke to him about it today; there seem to be differing visions for hoary [04:45] giggle.. neat! [04:45] jdub: what's yours? (the one you've announced on ubuntu-users@ with ETAs and stuff) [04:46] there are 1009 source packages in warty main [04:46] we've modified 308 of them [04:47] clobbering those changes with packages from unstable would produce an, err, undesirable result [04:48] elmo: would it be possible to rig something up to let unmodified packages flow in from sid, while not clobbering modified ones? [04:48] well, I can easily enough not sync if there's an 'unbutu' marker in the version number in warty [04:48] pittping [04:48] then we could work through those by hand [04:48] ugh [04:48] daniels: dude, your keyboard handling skills suck tonight [04:48] pitti's asleep [04:48] it's like 5am in Germany :) [04:48] hmm, and 4 am here.. sleep might be a plan soon [04:49] elmo: wow, way to make your point, dude :P === daniels considers a base-config upload with Canonical Standard Time (UTC+10) as the default TZ. [04:49] mdz: sure, no prob.. anyway, given the differing visions, shall I mail you, jeff, mark? [04:49] (think jeff's crashed now) [04:49] daniels: ? [04:49] oh, I thought he was still here [04:50] elmo: the 5am thing [04:50] it's like noon in jdub-land, isn't it? [04:50] wtf?? [04:50] To: apache@packages.debian.org [04:50] Subject: Andy Rourke (Smiths Bass Player) replaces Mani (Primal Scream) [04:50] and, sure enough, it's about andy rourke replacing mani [04:51] * jdub feels morbid, migrates to pillowful working area [04:51] mdz: mmm, thogh feeling uwjwell [04:51] mdz: 1251 [04:51] jdub: ah [04:51] mdz: so, just to make sure I'm not entirely on crack [04:51] mdz: ppp_on-boot_dsl is going into pppoeconf with a different name, no? [04:51] ppp_on_boot.dsl, even [04:51] daniels: s/pppoeconf/ppp/ [04:51] mdz: glad I asked [04:52] I thought I said that in the bug [04:57] mmm.. 4am.. night all [05:02] elmo: night dude [05:02] hmm, my flatmate has discovered the maximum possible score on tony hawk 2 [05:03] heh, not that hard to hit, iirc ;) [05:05] well, maximum score for a single trick [05:05] which is just over the 76 million mark [05:06] (spiderman + perfect balance + grinding the wire above the bullring + unplugging your keyboard + leaving it all night + a note that says 'leave please! for science's sake!' == very thps2 score) [05:06] s/very/& high/ === lamont bets Kamion went to bed, yes? [05:23] yes [05:32] mdz: want to change the debzilla address right now? (bug 1328) [05:32] justdave: yes [05:32] debzilla@canonical.com? [05:32] justdave: debzilla@ubuntu.com [05:32] email prefs are already turned off [05:33] ok, done. [05:33] make sure the importer knows to log in with the new address and you're all set. [05:33] done here [05:33] I'll do a test run [05:34] _seems_ to work [05:34] but I'm not sure there was anything new that represented a good test case [05:35] looks good though [05:42] hmm [05:43] who here has a shinybook? [05:58] define shiny :) [06:09] shinybook -> powerbooks === lamont takes polishing compound to his vaio === daniels waits for his pay, so he can get an X40. === lamont sends of his initial writeup of 'so you want to bootstrap an architecture' [06:11] heh [06:12] I have a Pismo here... [06:12] it's not exactly shiny [06:12] :) === doko [doko@dsl-084-057-004-250.arcor-ip.net] has joined #ubuntu-devel [06:42] morning guys === lamont sleeps. [06:49] night fabbione [06:51] ehe [06:51] night lamont [07:05] night lamont, morning fabbione [07:05] fabbione: i think the ati powerpc bios thing is an upstream bug [07:05] that's the reason why I wanted more testing :) [07:07] well i0t got some testing [07:07] it crashed in several machines [07:09] this is hoary stuff [07:09] X is now uber deep freeze [07:09] also.. apparently someone found a combinantion of nv + xv working [07:09] but if we go for that solution we need to roll back the nv driver to the one in -6 [07:09] that will not work for people that needs X.org driver [07:09] and the other way around [07:09] get the poiny [07:09] point even [07:14] yeah, there's no way we can do r4xx stuff for warty now [07:14] but if the bios detection is arse upstream on powerpc, i want to know that now :) [07:14] yeah [07:15] i think having the support xorg gives us is more importand than xv, though [07:15] remember, it's not terribly long until we'll have xorg [07:15] which reminds me -- early or late november for the sprint? [07:16] i think early is good, so we can get it done as quickly as possible [07:18] daniels: fine for me [07:18] did you mail Mark? [07:19] not yet, but i'll do that now === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [07:22] also because there is just a month or so left [07:22] so it's better to get organized [07:24] anyway Mark also said that we shouldn't spend too much time on X anymore since it is "stable" [07:24] and focus on all the other bugs [07:24] Morning guys! [07:24] hey pitti [07:26] fabbione: yeah, I'm trying to get off X [07:26] fabbione: working on Gimp now, for example [07:27] <- raidtools2 [07:34] fabbione: BTW, X on my iBook works again. Thanks for fixing the ati driver [07:35] pitti: no problem and thanks for reporting :-) [07:37] npmccallum: here? === m_tthew [matt@iorek.ice-nine.org] has joined #ubuntu-devel === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel === pitti [~martin@195.227.105.180] has joined #ubuntu-devel === rburton [~ross@82-44-126-41.cable.ubr03.croy.blueyonder.co.uk] has joined #ubuntu-devel === rburton [~ross@82-44-126-41.cable.ubr03.croy.blueyonder.co.uk] has joined #ubuntu-devel === azeem [~mbanck@socks-out.lrz-muenchen.de] has joined #ubuntu-devel === pitti [~martin@195.227.105.180] has joined #ubuntu-devel [11:08] jdub, mdz: ping [11:09] http://people.no-name-yet.com/~fabbione/raid_changelog [11:09] http://people.no-name-yet.com/patches/raidtools2.272864.dpatch [11:09] permission to upload raidtools2 to fix #1603 [11:09] diff and changelog at the above url [11:11] tested with all possible raids [11:11] there are no regressions here === seb128 [~seb128@ANancy-111-1-26-32.w81-51.abo.wanadoo.fr] has joined #ubuntu-devel [11:18] Hi seb128! [11:18] hello [11:18] seb128: the new gnome-vfs broke hotplug device handling :-) [11:18] pitti: og [11:18] oh [11:18] seb128: #1599, but I'm just fixing it [11:19] I've seen your mail to ettore, this guy works on gnome-vfs ? [11:19] seb128: I copied this from debian/copyright [11:19]