[12:00] <pitti> mdz: ^
[12:00] <pitti> mdz: but I think since we don't have NMUs in Ubuntu, this scheme should not clash with regular uploads
[12:00] <mdz> pitti: this is not an easy question
[12:01] <mdz> -2ubuntu0 << -2woody1, for example
[12:01] <mdz> which is what the Debian security team would use
[12:01] <pitti> mdz: right, but we don't use woody version numbers, do we?
[12:01] <pitti> mdz: well, _if_ we update a woody version, then we need to think about a different scheme
[12:01] <pitti> -2woody1ubuntu0.1
[12:02] <pitti> would be consistent, but ugly
[12:02] <pitti> but I think that won't happen very often
[12:04] <pitti> mdz: okay, I uploaded glibc (#2743) and imagemagick (#2769)
[12:07] <pitti> mdz: gee, we urgently need version (or at least distribution) tracking in our bts
[12:07] <mdz> pitti: you are learning the pain of security :-)
[12:07] <pitti> mdz: I left the bug open for now (normal/target Hoary) until I upload them to hoary
[12:08] <pitti> mdz: but I think I deal with Hoary when I fixed all other security bugs
[12:08] <pitti> mdz: that okay for you?
[12:10] <mdz> pitti: yes
[12:10] <mdz> pitti: I highly recommend keeping your own todo list
[12:10] <mdz> separate from bugzilla
[12:10] <pitti> mdz: in any case
[12:11] <mdz> pitti: fortunately sarge sorts below ubuntu :-)
[12:11] <pitti> mdz: and etch, too :-)
[12:12] <mdz> it should become part of our automated upgrade testing to ensure that this doesn't occur
[12:12] <pitti> mdz: and at the time etch is actually released, I'm grandfather
[12:12] <pitti> mdz: you mean it does not occur that our updates sort below the current Debian stable?
[12:12] <mdz> pitti: not only the current Debian stable, but anything that we support upgrades from
[12:13] <mdz> woody->warty is a supported upgrade path
[12:13] <mdz> so warty must be > woody
[12:13] <lamont_r> mdz: I thought sarge would do 1.sarge...
[12:13] <mdz> I believe we checked at one point
[12:13] <mdz> and all of the packages in warty/main had higher version numbers than woody
[12:13] <mdz> which is fortunate
[12:14] <lamont_r> mdz: that won't always be true, you know...
[12:14] <mdz> lamont_r: right, but we should ensure that it is true at release time
[12:14] <mdz> until we fix this awful version number problem once and for all
[12:15] <pitti> mdz: oh, gaim was missing
[12:16] <pitti> mdz: according to my log, I already uploaded gaim to the security queue, but it is not yet in the security packages file
[12:16] <lamont_r> mdz: Just thinking of the case where sarge happens to release just before the grumpy upstream version freeze, or some such.. :-(
[12:17] <pitti> mdz: I just checked all my logs, gaim is the only security upload I forgot about in above list
[12:17] <mdz> pitti: the security queue is different from the normal upload queue
[12:17] <mdz> pitti: packages do not automatically go into the archive via security
[12:17] <mdz> they must be processed by hand
[12:17] <pitti> mdz: yes, that's why I'm asking whether it is still present
[12:17] <mdz> gaim                        | 1:1.0.0-1ubuntu1.1    | source amd64 i386 powerpc     | 2 days old
[12:17] <pitti> mdz: or whether I somehow forgot to actually upload
[12:18] <pitti> ah, okay
[12:22] <pitti> mdz: I prepared the first USN for gs-common, of course the package list is still missing.
[12:22] <pitti> mdz: http://www.piware.de/usn-gs-common.txt , could you please have a short look? 
[12:22] <mdz> pitti: great
[12:22] <mdz> certainly
[12:22] <mdz> pitti: the date should be right-justified
[12:23] <mdz> but I think amber should do that for you now
[12:23] <pitti> mdz: oh, amber also creates this template?
[12:23] <pitti> mdz: I just copied the template from the website, where it is not justified at all
[12:23] <pitti> mdz: fixed
[12:24] <mdz> pitti: yes, amber will give you a template to be filled in
[12:24] <pitti> very nice
[12:24] <mdz> the website stuff is cut-and-waste from the emails I sent
[12:24] <pitti> I prepare the stuff for the other packages and send them out as soon as I get the amber mail
[12:25] <pitti> Probably not today any more, I'm exhausted and it's late
[12:25] <pitti> and libc6 will take a while to build (took 4 hours on my machine)
[12:25] <elmo> pitti: 30 mins in the DC :)
[12:26] <pitti> elmo: okay, okay, you have better machines to play with :-)
[12:26] <jbailey> pitti: How many passes is that?
[12:26] <elmo> hmm, actually, it's 30 i386, 43 powerpc, 48 amd64
[12:26] <pitti> jbailey: pardon?
[12:27] <elmo> pitti: how many different versions of glibc do you build for giggles
[12:27] <jbailey> pitti: processor/thread combinations.  In Debian on ia32, we usually have 3 (386-linuxthreads, 386-nptl, 686-nptl)
[12:27] <pitti> I just tried it on my desktop (686)
[12:28] <pitti> but I thought debuild would build them all?
[12:28] <jbailey> 4 hours just seems awful high.
[12:28] <pitti> I remember seeing much of the build log parts more than once
[12:28] <elmo> jbailey: it's stock Debian source, so it'll be whatever debian does for 386
[12:28] <elmo> s/stock/pretty-much \1/
[12:29] <pitti> jbailey: I don't have the fastest machine, though :_)
[12:31] <mdz> pitti: since we only upload source, it is important to test the binaries which are built by the buildds
[12:31] <mdz> at least one arch
[12:31] <jbailey> azeem: Debian GNU/Hurd takes about 4 times as long in my experience.  I usually leave 3 hours on my 1.8ghz p4.
[12:31] <pitti> mdz: that's why I compiled the stuff and tested it, and why it took so long until I actually asked for approval :-)
[12:31] <jbailey> (for linux)
[12:32] <mdz> pitti: hm?
[12:32] <jdub> yo jbailey 
[12:32] <mdz> pitti: the binaries built by the buildds
[12:32] <azeem> jbailey: it definetely wasn't 12 hours on my 1.8ghz p4
[12:32] <azeem> (on Hurd)
[12:32] <jbailey> g'morning Jeff. =)
[12:32] <pitti> mdz: ah, I see
[12:32] <azeem> more like 5 or 6
[12:32] <pitti> mdz: sorry, confused
[12:32] <jbailey> azeem: We do 3 passes on i386, though.  On the Hurd we only do one.
[12:32] <azeem> ah :)
[12:46] <jdub> mdz: how'd the meeting go? i fell asleep late last night putting pipka to bed
[12:46] <seb128> hello jdub 
[12:46] <jdub> pants off seb128 
[12:46] <seb128> :)
[12:46] <mdz> jdub: which one? CC?
[12:46] <jdub> yeah
[12:46] <mdz> it went pretty well, I was in and out a bit at the end
[12:46] <mdz> jdub: mako has transcript and is preparing a summary
[12:46] <jdub> cool
[12:50] <mako> jdub, mdz: am planning on finishing that one tonight
[01:05] <pitti> mdz: I have all USNs ready. I need some sleep, I will send them out tomorrow (when the amber mails hopefully arrived)
[01:05] <pitti> Good night everybody!
[01:07] <mdz> pitti: wait
[01:07] <pitti> mdz: still here
[01:07] <mdz> pitti: the amber mails will not arrive until I run amber
[01:07] <mdz> and the advisories should go out immediately after that
[01:07] <mdz> glibc does not seem to be built yet
[01:08] <jdub> daniels: around?
[01:08] <pitti> hmm, is there an approximate ETA?
[01:08] <pitti> mdz: I can stay up for another hour, if necessary
[01:12] <mdz> pitti: elmo is making some adjustments to amber
[01:12] <mdz> pitti: which ones are you ready to send out?
[01:12] <pitti> mdz: all but gaim
[01:12] <pitti> mdz: I'm currently typing gaim
[01:12] <mdz> pitti: ok. I have tested gaim and it works fine; I can't test MSN but I see no regressions in AIM/jabber
[01:13] <mdz> pitti: which ones are the simple temporary file stuff?
[01:13] <mdz> gs-common, glibc, gettext....?
[01:13] <pitti> mdz: gettext, glibc, gs-common, postgresql
[01:13] <pitti> mdz: do you have the two CC'ed mailing list adresses at hand?
[01:14] <mdz> pitti: gaim, gettext, gs-common, postgresql are built and ready
[01:14] <pitti> mdz: please give me some more minutes
[01:14] <mdz> glibc and imagemagick are not built yet
[01:14] <pitti> mdz: I need to change my subscription to bugtraq
[01:15] <pitti> mdz: I subscribed as martin@piware.de, but I need my canonical address
[01:15] <Kamion> mdz: I'm producing the required groff patch now, FYI
[01:15] <mdz> Kamion: thanks, keep pitti informed
[01:15] <mdz> pitti: bugtraq@securityfocus.com, full-disclosure@lists.netsys.com
[01:16] <pitti> mdz: do I need to subscribe full-disclosure to post to it?
[01:16] <mdz> pitti: I am not sure
[01:16] <mdz> pitti: you need to be subscribed anyway
[01:17] <pitti> mdz: okay
[01:17] <daniels> jdub: sup?
[01:18] <jdub> daniels: do you think composite-on-by-default is likely for hoary?
[01:19] <daniels> jdub: we can try it, but I'm uncomfortable doing it without a fair bit of testing
[01:19] <daniels> but if we test it in the hoary tree and it's not too bad, we can keep it on for main
[01:19] <jdub> tempting
[01:20] <jdub> though it's going to be different for different video cards, right?
[01:22] <daniels> yeah
[01:22] <daniels> due to render accel
[01:22] <daniels> i think radeon's is pretty solid, and that in the nvidia binary driver can be a bit odd sometimes
[01:22] <lifeless> is it easy to turn on/off ? I.e. a control panel ?
[01:22] <daniels> lifeless: Section "Extensions"\n\tOption\t"Composite"\t"Enabled"\nEndSection
[01:23] <daniels> it's not configurable on-the-fly
[01:23] <lifeless> ah. perhaps that should be a goal - an option for folk to switch, but htey have to log out and in
[01:23] <mdz> pitti: are you ready to publish?
[01:23] <mdz> pitti: elmo is finished and I am ready to amber
[01:23] <pitti> mdz: I don't get a confirmation email from netsys
[01:24] <pitti> mdz: I subscribed several minutes ago...
[01:24] <daniels> lifeless: mmm
[01:24] <pitti> mdz: otherwise, bugtraq subscription is okay, all USNs are ready
[01:24] <mdz> pitti: the important thing is that it goes to ubuntu-security-announce and the webiste
[01:24] <daniels> lifeless: i'll see how it holds up in hoary; if it's stable enough for what we're doing, great, enabled per default.  if not, it'll be disabled per default, and I'm not so sure I don't want a shiny-things button.
[01:24] <pitti> mdz: I can send the full-disclosure mail later if it has member-only posting 
[01:24] <daniels> s/don't want/want/
[01:25] <mdz> pitti: have you read the advisory guidelines I put in the wiki?
[01:25] <pitti> mdz: not yet
[01:25] <mdz> pitti: http://wiki.ubuntu.com/SecurityNotification
[01:26] <pitti> mdz: ah, I read this
[01:26] <mdz> pitti: for the temporary file vulnerabilities, the advisory should indicate with what privileges the overwriting takes place
[01:26] <mdz> if it is the privileges of a user invoking a program, or a system process, etc.
[01:26] <mdz> because this makes a big difference in the impact
[01:26] <elmo> hmm, lamont has the buildds set to take too much
[01:27] <pitti> mdz: hmm, I have to add this
[01:27] <elmo> well, maybe.. might not matter too much, even 10 packages isn't going take too long on average, and worst case will screw us even if we have less
[01:27] <mdz> pitti: let me know when you are ready for me to review the text
[01:28] <mdz> elmo: what are we going to do about getting new packages from sid into hoary/universe?
[01:28] <mdz> (rather than new versions)
[01:28] <mdz> or are you already doing that?
[01:28] <elmo> I'm going to deal with that in a bit.. also have to deal with removed packages, which is alittle more fun
[01:28] <elmo> since we have no record of where non-Debian stuff in universe came from (whee)
[01:28] <mdz> indeed
[01:29] <elmo> can someone prioitize merging/clearing-for-sync libldap2 and any other gnutls using libraries?
[01:29] <elmo> without that hoary's going to be severely broken dep-wise
[01:30] <elmo> mdz: btw, I assume I'm okay that people don't need to bother you/jeff for clear-to-sync mails?
[01:30] <Kamion> mdz: ok, uploaded to Debian, cced pitti on my mail to upstream
[01:31] <mdz> elmo: correct
[01:32] <mdz> elmo: dude, the merges are done
[01:33] <mdz> Keybuk said they were finished this afternoon
[01:34] <Kamion> I guess that would explain why I still have the majority of uploads to hoary :P
[01:39] <pitti> mdz: ready: http://www.piware.de/usn/
[01:39] <pitti> mdz: I hope my English is not too bad at this time
[01:39] <pitti> mdz: could you do some proofreading, please? For helping a USN author newbie? :-)
[01:41] <elmo> Recently, Trustix Secure Linux discussed a vulnerabily in the
[01:41] <elmo> discovered
[01:41] <elmo> vulnerability
[01:41] <elmo> (postgres)
[01:42] <elmo>  way, which allowed a symlink attack for overwriting
[01:42] <pitti> elmo: thanks, fixed
[01:42] <elmo> I'd say: way which allowed a symlink attack to overwrite
[01:42] <elmo> personally, but that's possibly just me
[01:42] <Kamion> agreed
[01:42] <pitti> elmo: I take the word of a native speaker
[01:43] <elmo> same for gs-common then
[01:44] <pitti> elmo: fixed both errors in all USNs
[01:44] <pitti> and uploaded again
[01:44] <elmo> A buffer overflow and two remote crashes were recently discovered in
[01:44] <elmo> the MSN protocol handler.
[01:44] <elmo> I'd say "gaim's MSN protocol handler"
[01:44] <elmo> otherwise it sounds like MSN is buggy, IYSWIM
[01:45] <pitti> fixed
[01:45] <mdz> elmo: which is not altogether inaccurate, but irrelevant :-)
[01:45] <mdz> pitti: reading now
[01:45] <elmo> pitti: still misspelling vulnerability in postgres
[01:45] <mdz> pitti: have you verified the CVE names against the CVE website?
[01:45] <pitti> mdz: yes
[01:46] <pitti> elmo: spell fixed
[01:47] <elmo> "Since imagemagick might be used in custom printing systems," <- s/might/can/ is more idiomatic and solves the duplication in the next sentence
[01:47] <pitti> done
[01:50] <mdz> pitti: it is a good idea to run an aspell check on all of the text
[01:50] <pitti> will do next time
[01:51] <tseng> hiya
[01:51] <mdz> pitti: s/files of/files with the privileges of/
[01:51] <mdz> pitti: in at least gettext
[01:52] <mdz> the files would not need to belong to that user, only be writable by them
[01:52] <mdz> pitti: also s/overwrite/create or overwrite/
[01:55] <pitti> mdz: fixed
[01:56] <tseng> cool, jimmac is using ubuntu
[01:56] <mdz> pitti: looks good
[01:57] <mdz> pitti: ready to release?
[01:57] <pitti> thumbs up! I even got my f-d subscription
[01:57] <elmo> hey, that's a point
[01:57] <elmo> we need a "who uses ubuntu" section on the website
[01:57] <jdub> oh dudes
[01:57] <jdub> http://primates.ximian.com/~jimmac/blog/Misc/Ubuntu
[01:57] <jdub> ^ jimmac likes Ubuntu on his Mac
[01:57] <tseng> oh dudes, i just said that !
[01:57] <tseng> :P
[01:57] <jdub> heh
[01:57] <mdz> who is jimmac?
[01:57] <daniels> elmo: that'd be easy
[01:57] <mdz> I assume he works for ximian
[01:57] <daniels> elmo: 'all the cool kids'
[01:57] <jdub> mdz: jakub steiner, gnome icons man
[01:58] <daniels> mdz: yeah, he's one of their crazy art guys
[01:58] <mdz> doh
[01:58] <mdz> elmo: amber blew up
[01:58] <jdub> and ximian/novell art biatch
[01:58] <elmo> sweet
[01:58] <mdz> elmo: jabbered you the output
[01:58] <elmo> no you haven't?
[01:58] <mdz> like hell I haven't
[01:59] <elmo> have NOT
[01:59] <mdz> sent again?
[01:59] <mdz> perhaps you are not logged into jabber :-P
[01:59] <elmo> I so am dude
[01:59] <mdz> ooh, jabber is broken
[01:59] <elmo> geez.. first he insists he's sent me mail.. then he insists he's sent me jabbers..
[01:59] <elmo> ;-)
[02:00] <mdz> I disconnected from jabber and tried to reconnect, and it hangs
[02:01] <jdub> haha
[02:01] <jdub> http://davyd.ucc.asn.au/images/battstat-preferences-cleaner.png
[02:02] <jdub> davyd's latest sshot has 'sudo suspend' ;)
[02:02] <jdub> mdz: btw, why did we go for "add a user line to sudoers" instead of a wheel-group equivalent set up?
[02:03] <mdz> jdub: no one even mentioned the possibility, including you :-P
[02:03] <mdz> could very well be simpler to do it that way
[02:03] <tseng> %wheel  ALL=(ALL)   ALL
[02:03] <jdub> mdz: btw, http://cipherfunk.org/diary/archives/monthly/2004-10.html#e2004-10-26T18_44_02.htm
[02:04] <jdub> mdz: i thought of it a few times, but stupidly didn't mention it, because i thought it had been passed over.
[02:04] <jdub> mdz: fwiw, that's how OS X does it
[02:04] <jdub> mdz: would it be a PITA to transition to it?
[02:04] <mdz> jdub: not for new installs
[02:04] <tseng> http://ronald.bitfreak.net/images/g-v-a.png <- whats different in this shot?
[02:05] <jdub> i guess it wouldn't for upgrades either; you'd only have to add one line
[02:05] <jdub> and existing users already have sudo love
[02:05] <jdub> tseng: dunno
[02:05] <tseng> looks the same old to me =/
[02:05] <jdub> tseng: although it has been rewritten
[02:05] <tseng> yes.
[02:06] <jdub> mdz: interesting patches
[02:07] <jdub> (paul is 'ultrafunk' aroudn irc)
[02:08] <jdub> you know
[02:08] <jdub> i was thinking last night
[02:08] <mdz> jdub: what did he have on the laptop before?
[02:09] <jdub> probably fedora
[02:09] <Kamion> mdz: I saw it mentioned a number of times, actually
[02:09] <jdub> we make a lot of decisions on irc
[02:09] <jdub> and talk about interesting stuff here
[02:09] <Kamion> mdz: always when more important stuff was happening, though ...
[02:09] <jdub> while some people may only be able to track ubuntu-devel
[02:09] <mdz> Kamion: I remember it was mentioned late in the release cycle
[02:09] <jdub> i want to make more of an effort to use that list
[02:09] <mdz> but not back in Oxford when we made the change
[02:09] <Kamion> mdz: *nod*
[02:10] <Kamion> mdz: I think it'd be good for new installs, but I was a bit scared to use "wheel"; that group has a lot of baggage
[02:10] <mdz> Kamion: agree
[02:10] <Kamion> people rather expect it to be gid 0, too
[02:10] <elmo> god damn it, I suck.
[02:10] <mdz> pitti: still with us?
[02:10] <pitti> mdz: yes
[02:10] <Kamion> adding the initial user to gid 0 would suck :-)
[02:11] <mdz> pitti: minor train wreck in generating the advisory template
[02:11] <Kamion> anyway, this alleged going-to-bed thing ...
[02:11] <jdub> adduser jdub sudo ;)
[02:11] <daniels> Kamion: ... is for the weak!
[02:11] <pitti> mdz: in the meantime I created all mails and put them into my outgoing queue; so I just need to add the package lists
[02:12] <Kamion> I'll be out for a fair bit of tomorrow; picking up the girlfriend from visiting family an hour or so's drive away
[02:12] <mdz> Kamion: ok, enjoy
[02:12] <daniels> enjoy :)
[02:12] <Kamion> will catch up as per usual :)
[02:15] <elmo> lol
[02:15] <elmo> katie@jackass:/srv/ftp.no-name-yet.com/queue/accepted $ amber 3-1 gs-common_0.3.6ubuntu1.1*.changes -n | less
[02:15] <elmo> amber is only useful for members of the security team, you tramp.
[02:15] <Kamion> :-)
[02:15] <Kamion> you can smell elmo-written scripts a mile away
[02:16] <daniels> heh
[02:18] <jdub> they smell like schhhnnaaaakke
[02:34] <pitti> Hi sjoerd! nightshift, too?
[02:34] <mdz> pitti: you should have an email with the gs-common template now
[02:34] <pitti> mdz: not yet
[02:35] <elmo> Oct 27 01:34:10 fiordland postfix/smtp[10351] : D3041B68002: to=<mdz@alcor.net>, orig_to=<security@ubuntulinux.org>, relay=redir-mail-telehouse1.gandi.net
[02:35] <elmo> [217.70.180.1] , delay=2, status=sent (250 Ok: queued as ED36E445C1D)
[02:35] <elmo> Oct 27 01:34:39 fiordland postfix/smtp[10352] : D3041B68002: to=<martin@piware.de>, orig_to=<security@ubuntulinux.org>, relay=mail.piware.de[213.9.79.162] 
[02:35] <elmo> , delay=31, status=sent (250 OK id=1CMblu-0004Rv-VO)
[02:35] <mdz> it liked me better
[02:35] <pitti> mdz: now it's there
[02:36] <pitti> probably took a while to get though SA, procmail, and the like
[02:36] <mdz> pitti: gettext inbound
[02:36] <mdz> and postgresql
[02:37] <mdz> imagemagick
[02:37] <mdz> and gaim
[02:37] <elmo> hmm, glibc's taking too long, lemme check on it
[02:38] <pitti> mdz: the template shows all generated debs; shall I cut that down to the really affected ones?
[02:38] <elmo> ok, we need lamont for glibc
[02:38] <mdz> hmm, he was just here
[02:38] <elmo> the buildd mail config is broken for the powerpc and amd64 buildds it built on
[02:38] <mdz> GAH
[02:38] <mdz> can we retry it and hope it gets a different one?
[02:38] <elmo> lol
[02:39] <mdz> sorry, channeling lamont there
[02:39] <mdz> I think I am going to need to go buy some sandbags and build a dike
[02:39] <mdz> before I drown
[02:40] <pitti> mdz: ready to fire the first USN (gs-common)
[02:41] <mdz> pitti: whenever you are ready, or if you want me to look at it first, let me kno
[02:41] <mdz> know
[02:41] <pitti> mdz: it's just the known notice with the package list appended
[02:41] <pitti> to: u-s-a, cc: full disclosure, bugtraq, signed
[02:41] <mdz> pitti: go ahead
[02:44] <pitti> mdz: for gettext, only the really affected "gettext" or shall I include all other packages (gettext-base etc.) as well?
[02:44] <mdz> pitti: if you know that the bug is only in one binary package, you can list only that one
[02:45] <pitti> mdz: as I already did, okay
[02:45] <pitti> mdz: so I can also remove the other packages/md5sums/etc?
[02:45] <mdz> pitti: leavae the md5sums
[02:45] <mdz> leave
[02:45] <pitti> okay
[02:46] <pitti> gettext is away
[02:46] <pitti> I think I should wait for the arrival of gs-common before sending further mails...
[02:46] <mdz> I approved the list messages
[02:46] <pitti> argh, why does my post to u-s-a await moderator approval???
[02:46] <mdz> pitti: every post is moderated
[02:47] <pitti> ah, okay. So no error on my side
[02:50] <mdz> pitti: trying to reach lamont
[02:51] <mdz> he is enroute
[02:51] <pitti> ah, the first two mails came through :-)
[02:53] <mdz> pitti: :-)
[02:53] <pitti> mdz: I sent the three other USNs (all except libc)
[02:54] <mdz> pitti: do you have privileges to add them to the website?
[02:54] <pitti> mdz: time to try out my password :-)
[02:57] <elmo> you guys don't need me for anything else do you?
[02:57] <pitti> elmo: thanks a lot for your help!
[02:57] <pitti> mdz: it seems as if I could edit the page
[02:57] <mdz> elmo: all the ambering has worked fine, so I don't anticipate a problem with glibc, thanks
[02:58] <elmo> well, mobile's on anyway, if something comes up.. 
[02:58] <elmo> night all
[02:58] <jdub> night elmo
[02:58] <lamont_r> elmo
[02:58] <lamont_r> still here?
[02:59] <pitti> mdz: when I click on "add error reference" on http://www.ubuntulinux.org/support/documentation/usn/, I get a permission denied
[02:59] <pitti> mdz: is that the right way to add another USN?
[02:59] <elmo> lamont: wha?
[03:00] <lamont_r> mdz said I needed to poke a couple buildd's... wondering which machines...
[03:00] <elmo> whatever built glibc
[03:00] <elmo> adare and ..
[03:00] <lamont_r> right
[03:00] <lamont_r> ah, yes
[03:00] <elmo> yellow
[03:01] <elmo> the glibc build log bounced, AFAICT
[03:01] <lamont_r> yes
[03:01] <elmo> do I want to know why, or will it make me cry?
[03:01] <elmo> the size limit seems to be set to 0..
[03:01] <lamont_r> message_size_limit, iirc
[03:01] <mdz> pitti: yes, that's right
[03:01] <mdz> pitti: you must not have permission
[03:02] <lamont_r> dunno - I'll get 'em signed/uploaded, and triage/fix/email you when I get home
[03:02] <pitti> mdz: can I get it?
[03:02] <elmo> lamont: k
[03:02] <mdz> pitti: I don't know how to do that, but I'll look
[03:03] <mdz> website....so....slomw
[03:03] <mdz> slow
[03:04] <mdz> pitti: I don't see a way to do it; I may not have permission
[03:05] <pitti> mdz: hmm; Alexander Limi and Jane should know how?
[03:05] <pitti> mdz: maybe we can sort that out tomorrow and you copy the USNs to the page?
[03:05] <mdz> pitti: yes, I will need to do that
[03:06] <pitti> mdz: sorry :-/ I will ask somebody tomorrow to give me permissions
[03:06] <lamont_r> pitti: glibc failed on amd64
[03:06] <pitti> damn
[03:06] <lamont_r> amd64 not in arch list. :-(
[03:06] <lamont_r> missed it by that much...
[03:06] <pitti> lamont_r: I only changed a script, after all...
[03:07] <elmo> so, err, yeah, how did it build last time?
[03:07] <pitti> lamont_r: does the Debian version fail as well?
[03:07] <lamont_r> could, dunno
[03:08] <lamont_r> elmo: glibc does absolutely insane and stupid things with it's control file, then patches both versions, so it doesn't patch it during build.
[03:08] <lamont_r> istr adding a touch to debian/rules for warty...
[03:08] <lamont_r> pitti: bbiab
[03:10] <mdz> wait, what?
[03:13] <mdz> does someone have the failed logs in hand?
[03:13] <mdz> and is that same someone working to fix the problem?
[03:15] <elmo> http://people.ubuntu.com/~james/glibc_2.3.2.ds1-13ubuntu2.1_20041026-2338.gz
[03:17] <elmo> I'm not working on it tho
[03:17] <mdz> pitti: all USNs should be published on the website now
[03:18] <pitti> mdz: nice, thanks
[03:19] <pitti> mdz: you did not include the pacakge md5sum lists, as you did for your first two USNs?
[03:19] <mdz> pitti: yes, I will remove them from the first two as well
[03:19] <jdub> mdz: btw, you don't have to turn emergency moderation on, or list yourself as admin and moderator
[03:19] <pitti> mdz: probably nicer
[03:20] <mdz> jdub: the other half of that sentence would have told me the alternative method to get what I want :-)
[03:20] <jdub> mdz: it was already set up correctly
[03:20] <mdz> jdub: it was set up for every message to be moderated?
[03:20] <jdub> yes
[03:20] <jdub> new members are moderated by default
[03:20] <jdub> on that list
[03:20] <mdz> new members?
[03:21] <jdub> subscribers
[03:21] <jdub> "By default, should new list member postings be moderated?" -> YES
[03:21] <mdz> where is that set?
[03:21] <jdub> sender filters page under privacy
[03:22] <mdz> first place I would have looked
[03:22] <mdz> NOT
[03:22] <mdz> what's a "new list member"?
[03:23] <mdz> ah, I see
[03:23] <mdz> so when new users subscribe, they get their individual moderator flag checked
[03:23] <mdz> I think I feel better with it being globally locked down
[03:26] <pitti> mdz: interesting post on u-users: sb asked how to verify GPG keys on the u-s-a list
[03:26] <pitti> mdz: shall we put the fingerprints on the website?
[03:26] <mdz> pitti: the same way to verify any other GPG key
[03:26] <mdz> through the web of trust
[03:26] <pitti> mdz: still it might be a good idea to have the fingerprint on a website
[03:26] <mdz> it only encourages bad habits, really
[03:27] <mdz> but we can do that
[03:28] <pitti> lamont_r: already back?
[03:31] <pitti> mdz: do you think the glibc issue will resolve in the next minutes? Or can we continue tomorrow, or can you send the announcement?
[03:31] <mdz> lamont_r: ?
[03:31] <pitti> lamont_r: ah, welcome back :-)
[03:32] <lamont_r> so, how do I get X to not go to sleep, never to return, every time I forget and close the lid on the laptop, huh?
[03:32] <mdz> lamont_r: purge acpid?
[03:32] <lamont_r> done
[03:32] <lamont_r> thanks
[03:32] <lamont_r> did I miss anything worth reading the log file for ?
[03:33] <pitti> lamont_r: not really
[03:33] <lamont_r> ok - I'll see it when I get home
[03:33] <mdz> lamont_r: you missed me losing my mind over how we managed to release warty with a glibc that FTBFS
[03:33] <lamont_r> which'll be relatively soon
[03:33] <lamont_r> mdz: I swear I didn't cheat on that one.
[03:33] <pitti> lamont_r: any ETA? it's 3:34am and I can hardly keep my eyes open
[03:33] <mdz> lamont_r: so you think it's a timing thing?
[03:34] <mdz> pitti: go ahead and sleep, I will take care of it
[03:34] <mdz> or it can wait until tomorrow
[03:34] <mdz> pitti: thanks for staying up
[03:34] <pitti> mdz: okay, I send you the announcement
[03:34] <pitti> mdz: thanks a lot for mentoring me!
[03:34] <lamont_r> mdz: glibc has traditionally been victim to that.
[03:34] <mdz> lamont_r: we need for warty-security build failures to go to security@
[03:34] <lamont_r> note: never patch both the source and target in the diff.gz, unless you deal with it somehow in rules...
[03:35] <pitti> mdz: USN sent to you
[03:35] <pitti> thanks everybody and good night!
[03:35] <mdz> pitti: night
[03:35] <mdz> lamont_r: can you fix it up with an upload to warty-security?
[03:36] <lamont_r> mdz: meaning fix and re-upload source, I assume?
[03:36] <mdz> lamont_r: correct
[03:36] <lamont_r> wilco
[03:36] <mdz> thanks
[03:37] <lamont_r> np.  car about done here, then will be home relatively soon and deal with it.
[04:05] <hornbeck> I am liking the new wiki :-)
[04:11] <jdub> mdz!!!
[04:11] <mdz> jdub???
[04:11] <jdub> there is an optimisation that we MUST HAVE in hoary!!!
[04:11] <mdz> oh really???
[04:12] <jdub> yes!
[04:12] <jdub> !!
[04:12] <jdub> it is crucial for enterprise adoption!!
[04:12] <jdub> http://tech9.net/rml/log/2004102601
[04:13] <mdz> ...
[04:13] <tseng> yeah, meh
[04:14] <mdz> jdub: we should add a shutdown script which executes sync(1) in a shell loop 100,000 times
[04:14] <mdz> then replace our sync(1) with this one and claim a performance benefit
[04:14] <hornbeck> jdub: that is nice
[04:15] <mdz> d00d3rz 1 r3c0mpi1e/> w1th -p1p3 4n/> 1t /2u/\/z 50kx f4st3r!!!11
[04:25] <jdub> mdz: that MSI bug seems to suggest that it was on and off during warty devel
[04:26] <mdz> jdub: yeah, it was supposed to end in the 'off' state
[04:26] <jdub> i don't want to build kernels anymore :-)
[04:26] <mdz> jdub: if you have a -15 or lower .deb around, you can try that
[04:26] <jdub> hrm
[04:26] <mdz> maybe a preview CD
[04:26] <mdz> the regression happened oct 12
[04:27] <mdz> snapshot.ubuntu.com? :-)
[04:27] <jdub> heh
[04:28] <mdz> fuck, raining again
[04:33] <lamont> home, even
[04:36] <lamont> Rejected: gtk2-engines-smooth_0.5.8-1_powerpc.deb: old version (2.8.1-0ubuntu1) in hoary >= new version (0.5.8-1) targeted at hoary.
[04:42] <mdz> lamont: what's the word on glibc?
[04:44] <lamont> mdz: the word is that I'm home and ignoring irc while I fix it. :-)
[04:44] <mdz> thanks
[04:51] <hornbeck> how would you make a code block in reStructuredText?
[04:52] <hornbeck> nevermind
[04:53] <mdz> hornbeck: you are actually using it?
[04:53] <mdz> I found it unusable
[04:56] <hornbeck> what reStructuredText
[04:56] <hornbeck> it is great
[04:56] <hornbeck> like butter on your breakfast toast
[05:04] <mdz> the first thing I tried to write, it blew up because the header markup wasn't exactly the same number of characters as the heading text
[05:04] <hornbeck> yeah, I have some problems with little lame things
[05:05] <hornbeck> like not having returns in the right places
[05:05] <hornbeck> man building beagle is a pain in the ass sometimes
[05:06] <jdub> mdz: sabdfl reckons we'll standardise on moin-style markup
[05:06] <mdz> jdub: me too
[05:06] <hornbeck> really
[05:06] <hornbeck> so I am learning this for nothing :-(
[05:06] <hornbeck> oh well
[05:06] <mdz> I can't believe you don't hate it :-)
[05:06] <hornbeck> I don't hate many things
[05:06] <hornbeck> I can find fun in pooping outside
[05:07] <lamont> mdz: have me remind you to beat up the person who added amd64 support 99% of the way, eh?
[05:07] <lamont> source upload should happen in about 45 minutes, give or take
[05:09] <hornbeck> is the universe repos down right now
[05:09] <hornbeck> my apt-get just died
[05:12] <lamont> hornbeck: new day begins. :-)
[05:12] <lamont> "days" are 30 minutes long...
[05:12] <lamont> so the odds of catching a Packages.gz update or mirror pulse are greater.
[05:12] <hornbeck> damn
[05:12] <hornbeck> I am trying to write a evolution-sharp doc, and I need all the deps to make sure I get them all
[05:12] <lamont> day begins at :03 and :33, pulse shortly thereafter
[05:13] <hornbeck> guess it will wait till tomorrow
[05:13] <hornbeck> oh so it will be back in a few?
[05:17] <hornbeck> ok, I can get somethings from universe but alot of xml files are gone?
[05:17] <hornbeck> did some stuff get taken out?
[05:18] <hornbeck> Err http://archive.ubuntu.com warty/universe libxml-namespacesupport-perl 1.08-3  404 Not Found
[05:18] <hornbeck> any ideas?
[05:18] <hornbeck> mdz?
[05:20] <mdz> hornbeck: hmmm!
[05:20] <hornbeck> that package and libxml-sax-perl
[05:21] <hornbeck> both used to be there
[05:21] <mdz> it was moved, but not intentionally
[05:21] <mdz> I think I see what happened
[05:21] <mdz> elmo: ^^^^
[05:21] <mdz> (when he wakes up)
[05:22] <hornbeck> they will be back correct?
[05:22] <hornbeck> cause evolution-sharp depends on them
[05:22] <hornbeck> :-)
[05:23] <mdz> hornbeck: as I said, it looks like it was unintentional
[05:23] <hornbeck> ok cool
[05:23] <mdz> I'm writing an email to elmo
[05:23] <hornbeck> thank you
[05:25] <mdz> if you notice anyone else encountering that, let them know that we're aware of it and working on the problem
[05:26] <hornbeck> mdz: ok, I am off to bed anyway but was working on one more doc before bed
[05:30] <lamont> mdz: if it helps, I have a never-break-my-archive (albeit crufty) script for mirroring
[05:31] <lamont> snags Packages.gz/Sources.gz, parses them. fetches files, installs Packages.gz Sources.gz, fills the morgue (since bandwidth hurts me.)
[05:32] <lamont> mdz: so does squirrelmail make baby jesus cry?
[05:34] <jdub> yes
[05:34] <jdub> plus it uses baby-jesus-baiting php4-imap
[05:35] <lamont> any sane alternatives?
[05:36] <jdub> i haven't really found one that made me comfortable
[05:37] <jdub> atm i'm using sqwebmail at home, which has its own faults
[05:37] <lamont> that was what led me to installing squirrelmail on a non-i386 box
[05:38] <lamont> glibc (if/when it decided to rebuild debian/control) lacked amd64 mention in control.in/libc6 :-(
[05:38] <lamont> not sure where I got the amd64 patch when I applied
[05:38] <lamont> it
[06:29] <fabbione> morning guys
[06:38] <lamont> morning fabbione
[06:42] <fabbione> lamont: how busy are adaire and yellow?
[06:42] <fabbione> i guess all the buildd are pumping 100% after the sync
[06:42] <lamont> fabbione: building hoary, basically
[06:42] <fabbione> ok
[06:42] <fabbione> i will soon need to start building X.org
[06:42] <fabbione> it compiles on i386 already
[06:42] <lamont> ah, ok.  happy, happy, joy, joy
[06:42] <fabbione> lamont: i won't be that happy
[06:42] <fabbione> it is still monolithic
[06:42] <lamont> yeah - but the breakup of the source can largely happen in parallel with people abusing the monolithic beast.
[06:42] <lamont> you're going to break it up so that it doesn't have circular build-deps, right?
[06:43] <lamont> my mirror is now out-of-date: 1353 files missing. :(
[06:44] <lamont> well, 1249 when I snap debian sources over.
[06:44] <fabbione> lamont: i already did basically :-)
[06:49] <jdub> morning Keybuk 
[06:49] <fabbione> and i also have much less compiler warnings
[06:49] <lamont> fabbione: yeah
[06:49] <jdub> lamont: i386 packages.gz for hoary looks fat! :)
[06:49] <fabbione> than in the monolithic
[06:49] <fabbione> lamont: but it won't be uploaded until hoary release date + 1
[06:49] <Keybuk> jdub: morning
[06:49] <fabbione> so everybody will live with monolithic for another release cycle
[06:49] <jdub> lots of ubuntu packages in hoary ;)
[06:49] <lamont> jdub: scary thing is that universe in hoary got bigger than warty
[06:49] <lamont> ditto main
[06:49] <lamont> and multiverse
[06:49] <jdub> doesn't that make sense?
[06:49] <Keybuk> jdub: I have 77 to upload at the moment, another 277-odd remaining :)
[06:49] <jdub> Keybuk: rad!
[06:49] <jdub> lamont: given that building the livecd is 'hard', and rsyncing is not entirely useful, what do you think is the best way of pulling regular livecds for testing?
[06:49] <jdub> maybe we should have a jigdo equivalent for livecds
[06:49] <jdub> that'd be rad
[06:49] <jdub> oh
[06:49] <jdub> no it wouldn't
[06:49] <jdub> it'd be pretty much the same as doing a livecd build anyway
[06:49] <fabbione> lamont: i will quite soon (a couple of weeks from now) need the tools to build the livecd
[06:49] <lamont> jdub: that'll be scary
[06:49] <lamont> fabbione: np
[06:49] <fabbione> lamont: to sync the X autoconfig stuff on both live and install
[06:50] <lamont> I should have the doc done by then, esp if someone picks on me
[06:51] <fabbione> lamont
[06:51] <fabbione> otherwise we can find something else.. hmmmm
[06:51] <fabbione> actually
[06:51] <fabbione> there is something more cool we can do...
[06:57] <fabbione> lamont: is there any simple way i can detect we are running on a LiveCD?
[06:57] <fabbione> like a specific installed file?
[06:57] <fabbione> that it will be there for sure 100%
[06:57] <fabbione> since people have been asking for autoreconfig in case of a video card change
[06:57] <fabbione> that can be done in a init script
[06:57] <lamont>  /MorphixCD/  iirc
[06:57] <fabbione> the same init script can check if we are running on a LiveCD
[06:57] <fabbione> and instead of prompting it could reconfigure X
[06:57] <fabbione> that will save me the time to build liveCD's
[06:57] <lamont> we may have to actually schedule said script to run, btw.
[06:57] <fabbione> good
[06:57] <fabbione> if it needs to be done maually even better
[06:57] <lamont> yeah - telling that you're LiveCD isn't all that hard.  If it is, we'll make it not-hard. :-)
[06:57] <fabbione> i can stick it in /usr/share/xorg/reconfigure_for_live_cd
[06:57] <fabbione> or something
[06:57] <fabbione> and than you can force it manually
[06:57] <lamont> elmo/thom aroudn?
[06:58] <lamont> pitti around
[06:58] <lamont> ?
[06:58] <jdub> 184 upgraded, 4 newly installed, 6 to remove and 0 not upgraded.
[06:58] <jdub> Need to get 69.6MB of archives.
[06:58] <jdub> After unpacking 7676kB disk space will be freed.
[06:58] <jdub> heh
[06:58] <jdub> The following packages will be REMOVED:
[06:58] <jdub>   alsa-base alsa-utils gdm python2.3-genetic ubuntu-base ubuntu-desktop
[06:58] <jdub> d'oh
[07:00] <Keybuk> jdub: what you upgrading to?
[07:00] <Keybuk> lamont: that would be a totally insane thing to do right now, surely? :o)
[07:00] <lamont> Keybuk: beyond doubt
[07:01] <Keybuk> some stuff is old, from warty; some stuff is building from unstable; and other stuff is sitting on rookery waiting for me to get my key in the keyring
[07:01] <Keybuk> lamont: what do you hope they do?
[07:01] <lamont> put the files I just uploaded in the right place.
[07:01] <jdub> yeah, just seeing the state of the archive
[07:01] <lamont> 75 seconds, and we'll know... :-(
[07:04] <Keybuk> jdub: there will be a few problems where stuff sync'd from unstable depends on a package not sync'd because we modified it
[07:05] <jdub> i imagine gdm is waiting in your queue
[07:06] <Keybuk> rookery scott% grep -c "^diff " results/gdm_2.6.0.4-1ubuntu1.dropped
[07:06] <Keybuk> 11
[07:06] <Keybuk> that'll need manual review
[07:08] <fabbione> Keybuk: if it's only question of signing and upload i can do that for you
[07:09] <fabbione> if you think that will speed up stuff around
[07:09] <Keybuk> thanks, but by the time I've gone through the ones with po-dropped I'll hopefully have over a hundred or more to play with and elmo will be awake ;p
[07:12] <lamont> jdub: as for status:
[07:12] <lamont> hoary.all.amd64:Total 1599 package(s) in state Needs-Build.
[07:12] <lamont> hoary.all.i386:Total 1047 package(s) in state Needs-Build.
[07:12] <lamont> hoary.all.powerpc:Total 1354 package(s) in state Needs-Build.
[07:13] <jdub> cool
[07:13] <fabbione> Keybuk: ok
[07:13] <Keybuk> http://primates.ximian.com/~jimmac/blog/Misc/Ubuntu
[07:14] <Keybuk> \o/
[07:14] <Keybuk> Ximian employee in "If somebody asks me what linux distro he should try on his mac, I'm resolutely recommending Ubuntu." shock.
[07:14] <jdub> innit great? :)
[07:14] <lamont> cool
[07:15] <lamont> now we just need to make hoary suck less than warty, eh?
[07:16] <mdz> Keybuk: morning
[07:16] <hornbeck> is there a bootstrap to use to setup a ubuntu chroot?
[07:16] <lamont> mdz: mcmurdo, royal, yellow.
[07:16] <lamont> hornbeck: start with debootstrap from warty
[07:17] <Keybuk> mdz: morning
[07:17] <lamont> debootstrap ... warty
[07:17] <hornbeck> great thanks
[07:18] <mdz> Keybuk: what's the status of your merge project?
[07:18] <lamont> Keybuk: did we ever decide on a binNMU standard?
[07:18] <lamont> version number, that is.
[07:18] <Keybuk> mdz: 77 done with nothing dropped, working on automating .po now so hopefully that'll give us a bunch more, then there will be a hundred or so with patches to manually review
[07:18] <mdz> Keybuk: done == uploaded?
[07:18] <Keybuk> lamont: no, need to get you, elmo, neuro, aj, kamion, etc. together and let you fight it out
[07:19] <Keybuk> mdz: not uploaded yet
[07:19] <mdz> Keybuk: do you have a list of those which will definitely require manual review?
[07:20] <Keybuk> mdz: not yet, because I hope to get the number down; it's at 251 at the moment
[07:21] <mdz> Keybuk: are you certain your key is not already in the keyring?
[07:21] <mdz> I was under the impression that it was
[07:22] <Keybuk> well, I did an upload the other day and it vanished into thin air
[07:22] <mdz> Keybuk: what package?
[07:23] <Keybuk> ubuntu-artwork, just before warty release
[07:25] <mdz> Keybuk: do you recall the version number?
[07:25] <Keybuk> 0.2.12-1
[07:26] <mdz> there's only one 0.2.12-1 in the log, and it was accepted
[07:27] <Keybuk> yes, but Colin signed that
[07:27] <Keybuk> he had to get me out of bed because mine didn't go through
[07:28] <Burgundavia> Who would I talk to about the expired cert on bugzilla?
[07:29] <mdz> Burgundavia: no one; it's already filed in bugzilla
[07:29] <Burgundavia> mdz: so I assume the powers that be are going to fix it?
[07:29] <mdz> https://bugzilla.ubuntu.com/show_bug.cgi?id=1278
[07:29] <mdz> Burgundavia: it's not a particularly high priority
[07:30] <Burgundavia> ok, just wondering
[07:30] <mdz> being able to spoof our bugzilla wouldn't be particularly rewarding :-)
[07:30] <Burgundavia> sort of ironic though
[07:34] <fabbione> GOOOOOOOOOOOOOOOOOOOOOOOODY
[07:34] <fabbione> patch forwarding is basically completed
[07:34] <fabbione> mdz: ^^^
[07:34] <fabbione> mdz: + we build on i386
[07:34] <lamont> Starting proxy server: 2004/10/26 23:32:36| WARNING cache_mem is larger than total disk cache space!
[07:34] <lamont> hrm...wonder how to bump the disk cache size..
[07:35] <mdz> lamont: cache_dir
[07:37] <fabbione> mdz: now we need to fix all the packaging part :-)))
[07:37] <fabbione> that it's going to be... hmm evil
[07:39] <fabbione> mdz: since we switched to monolithic i will NOT neeed root on amd64 or ppc
[07:39] <fabbione> mdz: all the build-deps should be there already
[07:39] <mdz> fabbione: what about for testing?
[07:41] <fabbione> mdz: even if i have root on our boxes, there is nobody in front of them to see what's happening on the screen
[07:42] <fabbione> i will need people with amd64 and ppc boxes to test
[07:42] <fabbione> there is really nothing i can automate myself there
[07:49] <Keybuk> zsh: command not found: msgmerge
[07:49] <Keybuk> gah!
[07:52] <Keybuk> anyone !elmo have root on rookery? :)
[07:55] <mdz> hom
[07:55] <mdz> thom
[08:05] <lamont> g'night all
[08:07] <doko> morning lamont
[08:08] <jdub> night lamont 
[08:13] <Keybuk> gah, I really cannot work this frickin' po file stuff out
[08:27] <Burgundavia> cya all
[08:27] <Keybuk> I really don't think msgmerge is going to work for this stuff
[08:43] <sid77> hi all
[08:58] <maswan> because:
[08:58] <maswan> elmo: @ERROR: max connections (25) reached - try again later
[08:58] <maswan> (this night)
[08:59] <fabbione> pitti: where is USN-4-1 ?
[09:00] <fabbione> ehm
[09:00] <fabbione> mdz: ^^
[09:08] <Keybuk> not released yet, I'd guess
[09:08] <Keybuk> I suspect they're doing the same as Debian and assigning the number at the start of the security procedure, not the actual release of the notice
[09:10] <fabbione> yeah
[09:12] <fabbione> doh!
[09:12] <fabbione> it was on ubuntu-users
[09:12] <Keybuk> so I've actually "fixed" the .po issue
[09:12] <Keybuk> I have some Python that strips all the context from a patch and rejiggles the line numbers
[09:12] <Keybuk> so it actually works
[09:12] <Keybuk> muahahahaha
[09:13] <Keybuk> screw you hippy po comments
[09:16] <Keybuk>         if line.startswith(" ") or \                 (line.startswith("-#") and not line.startswith("-#,")) or \                 (line.startswith("+#") and not line.startswith("+#,")):
[09:16] <Keybuk> those po file people are as magic-punctuation happy as tom lord!
[09:16] <Keybuk> -#, and +#, are a change to the bit of the .po file that says "fuzzy" or "c-comment", etc.
[09:16] <Keybuk> I got the impression people would be upset with me if I dropped those
[09:17] <fabbione> hey tfheen 
[09:17] <fabbione> welcome back
[09:17] <tfheen> hiya
[09:17] <fabbione> JEEEEEEEEEEEE
[09:17] <Keybuk> tfheen: vawad's *always* ill :D
[09:17] <fabbione> x.org is taking a long time to build
[09:17] <fabbione> even with ccache
[09:17] <tfheen> Keybuk: she's started to eat networking cards now.
[09:18] <Keybuk> nice
[09:18] <tfheen> so, anything fun happened while I was away?
[09:18] <fabbione> the build log for the normal part is at least 2MB bigger than Xfree86
[09:18] <tfheen> apart from Warty releasing, obviously.
[09:24] <sid77> lol
[09:24] <tfheen> fabbione: get an amd64 with 4G of RAM and crosscompile? ;)
[09:24] <fabbione> tfheen: welcome to send me one :)
[09:24] <tfheen> fabbione :)
[09:24] <fabbione> model name      : Intel(R) Pentium(R) 4 CPU 2.00GHz
[09:24] <fabbione> it's not like i have a slow machine
[09:24] <fabbione> with 1 GB of ram
[09:24] <Keybuk> dude, it's a P-4 ... that's about 200Mhz in *real* CPU terms <g>
[09:24] <fabbione> model name      : AMD Athlon(tm) XP 2100+
[09:24] <fabbione> i also have this one with 1 GB of RAM
[09:24] <fabbione> it still takes hell of a lot to compile compared to XF86
[09:24] <fabbione> 7684 -rw-r--r--    1 fabbione fabbione  7852788 Oct 21 12:12 make_world.build.log
[09:24] <fabbione> 9940 -rw-r--r--    1 fabbione fabbione 10161505 Oct 27 06:56 make_world.build.log
[09:25] <fabbione> 3020 -rw-r--r--    1 fabbione fabbione  3087544 Oct 21 12:19 make_world_dbg.build.log
[09:25] <fabbione> 4600 -rw-r--r--    1 fabbione fabbione  4694731 Oct 27 07:24 make_world_dbg.build.log
[09:27] <tfheen> fabbione: you're supposed to _like_ multi-MB build logs.
[09:29] <Keybuk> @@ -81,1 +72,1 @@
[09:29] <Keybuk> -msgstr "Debian installatieprogramma-modules worden geladen"
[09:29] <Keybuk> +msgstr "Debian installatiemodules worden geladen"
[09:29] <Keybuk> -- 
[09:29] <Keybuk> cute
[09:29] <Keybuk> seems to be working (that's a dropped Debian patch, for obvious reasons)
[09:30] <tfheen> Keybuk: you're merging d-i branding?
[09:30] <Keybuk> tfheen: merging debian and warty to make hoary
[09:33] <fabbione> tfheen: oh yeah
[09:43] <fabbione> and for last:
[09:43] <fabbione>  868 -rw-r--r--    1 fabbione fabbione   880853 Oct 21 12:22 make_install.log
[09:43] <fabbione> 1224 -rw-r--r--    1 fabbione fabbione  1245422 Oct 27 07:43 make_install.log
[09:43] <fabbione> more than 50% BIGGER!
[09:43] <fabbione> OH YEAH
[09:46] <Keybuk> is that X.org, or just daniels? :o)
[09:47] <fabbione> X.org :-)
[09:47] <fabbione> i think i need a bigger machine
[09:47] <fabbione> i am sure debian buildd will NOT like this
[09:48] <Keybuk> * Creating copy of base-files
[09:48] <Keybuk> * Considering base-files
[09:48] <Keybuk> ! 2 patch hunks dropped
[09:48] <Keybuk> * Building base-files_3.1ubuntu1.dsc
[09:48] <Keybuk> -- cool, most the drops are easy (that's our issue & issue.net change)
[09:57] <maswan> fabbione: how about an 8-way opteron with 32 gigs of ram?
[09:58] <maswan> fabbione: a bit too expensive to get donated though, even if I probably could borrow one for a couple of weeks. :)
[09:59] <fabbione> maswan: useless
[09:59] <fabbione> X build doesn't fork across processors
[09:59] <maswan> fabbione: single-cpu speed more relevant?
[09:59] <maswan> ah
[10:00] <fabbione> 32GB would be still good :-)
[10:00] <maswan> Well, you could start by fixing that then. :)
[10:01] <fabbione> maswan: send me 2x16GB stick for a Dell Optiplex GX260 :-)
[10:01] <fabbione> i won't mind that kind of donation ;)
[10:02] <maswan> fabbione: heh. well, sticking within reality, you could get an hp dl585 with 32 gigs of ram too, with only 4 cpus. :)
[10:02] <maswan> fabbione: but then, those aren't usually free either
[10:03] <fabbione> ENOMONEY
[10:03] <fabbione> maswan: i recently bought a house and will get soon the second boss (formely called wife)
[10:03] <Keybuk> "Dear Mark, For Christmas I would like..." :o)
[10:03] <fabbione> Keybuk: "Dear Santa^WMark" ;)
[10:03] <maswan> fabbione: see, Keybuk has the general idea of it :)
[10:04] <fabbione> the first step will be to get my bigger office up and running
[10:04] <fabbione> second step one 19" rack
[10:04] <fabbione> than i will ask Santa to fill it up :P
[10:04] <Keybuk> now's the time to decide between air mailing xmas cards, or just e-mailing; isn't it :p
[10:05] <maswan> fabbione: a 2.4GHz or so opteron should be among the faster cpus for gcc grunting too, I think.
[10:07] <jamesh> the new Athlon 64 4000+'s look nice too
[10:07] <jamesh> if you just want a single processor machine
[10:07] <maswan> well, yeah, but those only fits a few gigs of ram
[10:07] <maswan> of course, they are lots cheaper though
[10:07] <jamesh> how much ram do you want?
[10:08] <maswan> fabbione thought 32 gigs would be nice :)
[10:08] <fabbione> with 32GB i can mount /usr/src in ram
[10:08] <jamesh> you'd need 4 processors for that, right?
[10:08] <fabbione> and compile in it :-)
[10:08] <fabbione> nope
[10:08] <fabbione> one processor or 2 are more than enough
[10:08] <maswan> well, yeah, I don't know of any board with more than 8 slots/cpu
[10:08] <jamesh> what's the largest DIMM you can get?
[10:09] <jamesh> given that each CPU can handle 4 sticks of memory
[10:09] <maswan> 1GB in practice, 2GB in theory
[10:09] <fabbione> maswan: get a few e10k cpu boards :P
[10:09] <maswan> jamesh: I've seen 8
[10:09] <maswan> fabbione: ah, but those do not have very fast single-cpu performance. :)
[10:09] <jamesh> maswan: for a single CPU AMD64?
[10:09] <fabbione> maswan: i know :-)
[10:10] <fabbione> they are still sweet tho
[10:10] <maswan> jamesh: well, yeah, you could just run a dl585 with one cpu. :)
[10:10] <jamesh> maswan: I've seen 8 slot dual processor AMD64 boards
[10:10] <jamesh> maswan: but the second 4 slots could only be used if you plugged in two CPUs
[10:10] <maswan> jamesh: well, are those 8 slots in total or 8 slots per cpu?
[10:10] <maswan> ah, yeah.
[10:10] <jamesh> since the memory controllers are on the CPUs
[10:10] <maswan> that's the thing, the dl585 has 8 slots per cpu
[10:11] <maswan> hmm.. you should be able to get a p4/xeon board with 8 or perhaps even 16 slots and just populate that with 1 cpu.
[10:11] <maswan> I think. I haven't looked much at that though.
[10:13] <jamesh> either way, you're looking at a lot of money
[10:13] <maswan> yeah
[10:15] <jamesh> it might be more economical to buy a bunch of Raptor hard drives and set up a RAID array
[10:22] <seb128> morning
[10:34] <fabbione> +usr/X11R6/lib/libXaw.so.8.0
[10:42] <seb128> that's ok to start working/uploading in hoarty now ?
[10:42] <fabbione> dpkg-buildpackage -rfakeroot -uc -us -b
[10:42] <fabbione> dpkg-buildpackage: source package is xorg
[10:42] <fabbione> dpkg-buildpackage: source version is 6.8.1-0.0
[10:42] <fabbione> dpkg-buildpackage: source maintainer is Fabio M. Di Nitto <fabbione@fabbione.net>
[10:42] <fabbione> dpkg-buildpackage: host architecture is i386
[10:56] <pitti> Morning!
[10:57] <pitti> mdz: still awake?
[10:59] <seb128> hello pitti 
[10:59] <pitti> Hi seb128!
[11:00] <seb128> ok, so nobody knows if we can start working/uploading in hoary if we should better wait a bit ?
[11:02] <pitti> seb128: I already uploaded some stuff and was told that it was okay
[11:03] <seb128> yeah, I've seen the mails on hoary-changes, but I was not sure with the syncs
[11:03] <seb128> thanks pitti 
[11:03] <pitti> seb128: I still remember elmo saying that the syncs are finished (the automatic ones, that is)
[11:05] <seb128> cool, thanks
[11:15] <pitti> elmo: Morning! Since mdz is already sleeping, can you do the libc unleashing, too?
[11:16] <Keybuk> sorta
[11:16] <Keybuk> basically there are three sets of things to sync
[11:16] <Keybuk> 1. things that haven't changed in warty or debian -- easy, nothing to do (so done)
[11:17] <Keybuk> 2. things that changed in debian, but not warty -- pull from debian, and build (elmo's set this running, it may be done, it may not)
[11:17] <Keybuk> 3. things that changed in both debian and warty -- need to merge, I'm working on this at the moment
[11:33] <enrico> Hello.  Is someone around that is involved in the Wiki migration?
[11:35] <enrico> justdave: around?
[11:49] <robtaylor> pitti: what do reckon to mdz's chroot idea for #2758?
[11:50] <robtaylor> pitti: i tried restarting hotplug, and running the same hald test, but didnt seem to change anything
[11:51] <pitti> robtaylor: well, I thought hotplug was not the problem
[11:51] <pitti> robtaylor: you'd need to run hal not in a chroot, but in the normal file namespace
[11:51] <robtaylor> pitti: you think it's kernel level?
[11:51] <pitti> robtaylor: I did not extensively deal with the live cd, though
[11:52] <pitti> robtaylor: no idea, sorry; please ask amu, he should know much better about the live cd
[11:52] <robtaylor> pitti: yeah, it a bit weird, i'm testing now with a more normal usb key and seeing exatly the same symtoms :/
[11:52] <robtaylor> pitti: cool. what timezone is amu?
[11:54] <pitti> robtaylor: mine, i. e. UTC+2
[11:56] <robtaylor> pitti: ah cool
[12:19] <enrico> Who's involved in the wiki migration?  I'd need to get in contatct with them
[12:20] <seb128> why not saying here what you want ? perhaps somebody can help you ?
[12:20] <fabbione> ciao enrico 
[12:20] <enrico> fabbione: ciao!
[12:20] <fabbione> seb128: that sounds familiar :-)
[12:20] <seb128> :)
[12:20] <enrico> fine
[12:20] <sid77> ciao! (i'm italian too ;)
[12:20] <enrico> Appearently the pages were migrated
[12:20] <enrico> sid77: ciao!
[12:20] <sid77> lol
[12:20] <enrico> However, the old wiki has not been locked
[12:21] <enrico> I had been assured by mark (and consequently I reassured the ubuntu-doc list) that the old wiki would have been locked before transition, so that no modifications would have been lost
[12:21] <enrico> So, I'd like to ask what is the migration path.  People in the list are puzzled
[12:21] <Kamion> Keybuk: hm, msgmerge was working absolutely fine for me, seemed to be perfectly automatable in principle
[12:23] <migus> hi
[12:23] <enrico> Thing is, wiki is the way people in ubuntu-doc are making all the work now, and it's been changing without notices for a while
[12:23] <enrico> Someone starts being annoyed of having something different or unclear every time they want to do some work
[12:24] <Kamion> SteveA has been doing most of the migration I think; he was looking around last night for somebody with suitable admin privileges to lock the old wiki.
[12:24] <fabbione> hi sid
[12:28] <fabbione> sid77: please no public away messages
[12:36] <__daniel> hai
[12:40] <fabbione> hmm intersting :-)
[12:40] <fabbione> pciutils wants to kill ubuntu-base
[12:40] <pitti> sjoerd: did you ever try to build hal-0.4.0 in a pbuilder? My initial upload to Hoary FTBFS, I have to build-dep on libxml-parser-perl
[12:40] <fabbione> (bootstrapping hoary)
[12:41] <pitti> fabbione: I just tried to switch my apt sources to Hoary, but the package files are still empty. Do I sth. wrong?
[12:41] <Keybuk> Kamion: I just couldn't get it to do the right thing, ever
[12:41] <fabbione> pitti: they aren't here...
[12:41] <Keybuk> given "warty.po" and "debian.po", I couldn't get one that had the best of both
[12:41] <fabbione> pitti: debootstrap warty and then change apt lines
[12:41] <pitti> fabbione: you mean they are not empty for you? Then it must be a problem of my provider...
[12:41] <Kamion> Keybuk: really did work fine for me
[12:41] <pitti> fabbione: my provider has a transparent proxy, maybe this still offers me the empty files
[12:42] <pitti> fabbione: thanks
[12:42] <Keybuk> Kamion: when I did msgmerge warty.po debian.po; I got the new debian strings, but the warty translations
[12:42] <Kamion> Keybuk: like I say, you need to produce hoary.po first using debconf-updatepo or whatever
[12:42] <fabbione> pitti: i am using a local mirror
[12:42] <Kamion> Keybuk: I wouldn't do it that way ...
[12:42] <Keybuk> when I did msgmerge debian.po warty.po I got the debian translations, but not the new strnigs
[12:42] <Kamion> Keybuk: both those approaches are wrong
[12:42] <Keybuk> Kamion: doesn't work for ordinary .po
[12:42] <Kamion> can you leave everything with d-i branding to me then? I don't want to have to redo it all
[12:42] <Keybuk> you won't have to :)  I had a lot more luck with mutating the patch
[12:42] <elmo> GAR. stupid moin.
[12:43] <Kamion> ordinary .po there's usually some kind of Makefile target to do it
[12:43] <Kamion> Keybuk: uh
[12:43] <elmo> does anyone know how to get this stupid thing to read it's updated config file?
[12:43] <Keybuk> I wrote some code to write patches that could apply to .po files, without the # bits getting in the way :p
[12:43] <Kamion> Keybuk: anything I can review before upload, please?
[12:43] <Keybuk> Kamion: is cooking now
[12:43] <Kamion> those comments are often needed for actual real live translators
[12:43] <Keybuk> Kamion: sure, but you can favour one set over the other
[12:44] <Kamion> true ...
[12:44] <Keybuk> it keeps the #, one
[12:44] <Keybuk> because that's important
[12:44] <Kamion> there are others too
[12:44] <Keybuk> but #: doesn't even need to be in the patch
[12:44] <Kamion> #: should be handled by debconf-updatepo in the debian/po/ case
[12:45] <fabbione> ok guys.. good news
[12:45] <sid77> sorry
[12:45] <Keybuk> *shrug* this way works :)
[12:45] <fabbione> X.org compiles both on AMD64 and ppc
[12:45] <Kamion> Keybuk: like I say, can I please just have a look at one of them before upload?
[12:45] <fabbione> now it's only question of fixing a few tons of packaging problems
[12:45] <Keybuk> Kamion: dude, someone's going to look at *all* of them before upload
[12:45] <Keybuk> you get a source package, debian->hoary diff and "I dropped these" file
[12:45] <Kamion> Keybuk: you said you were mass-uploading, I believed you :-)
[12:46] <elmo> jdub: ?
[12:46] <Keybuk> Kamion: yeah, for the ones with no "dropped" I can interdiff debian->hoary and base->warty quickly <g>
[12:46] <Kamion> 14:37 < Kamion> Keybuk: are you going to be doing a mass upload of merges, or what?
[12:46] <Kamion> 14:38 < Keybuk> Kamion: I think the basic idea is to do the merges as automatic as possible
[12:46] <Kamion> 14:38 < Keybuk> for any that it works, I'll upload
[12:46] <Keybuk> note "works" :)
[12:46] <jdub> elmo: here
[12:47] <elmo> jdub: do you know how to get moin to reload it's config?
[12:47] <elmo> and/or where you involved in setting our moins up?
[12:48] <pitti> Is Hoary automatically synced to sid now every day?
[12:49] <jdub> elmo: i set up the original one on rince -> it should just take on the configuration when a new cgi is run
[12:50] <elmo> pitti: yeah, where it can be
[12:50] <elmo> jdub: that's what I thought, it doesn't appear to be, tho :/
[12:50] <pitti> elmo: nice, thanks
[12:53] <fabbione> ccachetop is cool :-)
[12:53] <enrico> elmo: restarting apache?
[12:54] <elmo> enrico: done that
[12:55] <Kamion> Keybuk: fair enough
[12:55] <enrico> elmo: still modifiable, though
[12:56] <Keybuk> Kamion: if there really is some msgmerge magic, that'd be cool -- but I really couldn't get it to produce sane output
[12:56] <Keybuk> it tended to favour one set of translations over another, completely
[12:56] <Kamion> Keybuk: note I've done main-menu, cdrom-checker, cdrom-detect, yaboot-installer since I last talked to you
[12:56] <Keybuk> and dropped others, with no way of actually telling it dropped them without resorting to diff
[12:56] <enrico> elmo: sure that the new config is correct?
[12:56] <Kamion> Keybuk: I used msgmerge successfully on five packages with complex merges
[12:57] <Kamion> the output was exactly what I wanted
[12:57] <sjoerd> pitti: no i didn't 
[12:57] <elmo> aha.  STUPID moin.
[12:57] <elmo> right, it's done
[12:57] <pitti> sjoerd: it is very likely that it FTBFS for Debian, too
[12:57] <Kamion> I gave you the procedure; it would have to be adapted for simple po/ as opposed to debian/po/, but almost all our branding is in debian/po/
[12:58] <Keybuk> Kamion: yeah, and I followed that, and got crap out of the other end
[12:58] <sjoerd> pitti: yeah probably.. will check 
[12:58] <enrico> elmo: works, cool, thanks!
[12:58] <Keybuk> entirely commented-out blocks and most of the warty changes dropped
[12:58] <Kamion> Keybuk: what you told me you tried just above was totally different, though?
[12:58] <Kamion> Keybuk: entirely commented-out blocks are expected and useful, they act as notes of the unbranded translations
[12:58] <sjoerd> pitti: i'm currently checking out pmount, are you planning to create a group in later debian packages or something like that
[12:58] <Kamion> Keybuk: sounds like you did it the wrong way round
[12:59] <Keybuk> Kamion: I'm applying the debian changes to warty, not the warty changes to debian
[12:59] <Kamion> Keybuk: me too
[12:59] <pitti> sjoerd: If Debian agrees to one, it'll be my pleasure :-)
[12:59] <pitti> sjoerd: plugdev would make Ubuntu and Debian totally compatible...
[12:59] <Kamion> 11:42 < Keybuk> Kamion: when I did msgmerge warty.po debian.po; I got the new debian strings, but the warty translations
[12:59] <Kamion> 11:42 < Keybuk> when I did msgmerge debian.po warty.po I got the debian translations, but not the new strnigs
[01:00] <Kamion> you *have* to do the update against new source lines before attempting to msgmerge
[01:00] <Keybuk> yeah, warty.po there has had debconf-updatepo run over it
[01:00] <Kamion> after patching the rest?
[01:00] <Keybuk> yup
[01:00] <Kamion> also, you need to use a compendium with pre-updatepo-warty.po
[01:01] <Kamion> the compendium is essential
[01:01] <Kamion> as far as I can tell
[01:01] <Keybuk> I couldn't get that to make any difference
[01:01] <Keybuk> with/without -C produced the same file
[01:01] <sjoerd> pitti: who should agree with that :)
[01:01] <Kamion> all I can say is "worked for me" :)
[01:01] <Keybuk> yeah, I'll beat daf up to find some better way later
[01:01] <Kamion> I had to do some additional branding, but i expected that
[01:02] <pitti> sjoerd: well, before introducing a new group, it should at least be discussed on d-devel, or not?
[01:02] <Keybuk> applying po-tailored patch hunks seems to work too
[01:02] <sjoerd> pitti: for something like that, yeah probably
[01:02] <pitti> sjoerd: because it has wider use; e. g. libgphoto2 could use it, too
[01:02] <Keybuk> (after all, it has the ultimate same effect, "replace this msgstr with this one")
[01:02] <sjoerd> pitti: be my guest too propose it then 
[01:02] <sjoerd> pitti: libgphoto2 uses camera currently....
[01:03] <pitti> sjoerd: okay; I can explain the whole mechanism in Ubuntu and ask for opinions
[01:03] <sjoerd> pitti: yeah would be nice
[01:03] <pitti> sjoerd: camera is certainly okay, but does Debian really want camera, usbdrive, pcmciadrive, etc.
[01:03] <pitti> sjoerd: adding to my TODO list
[01:03] <sjoerd> pitti: dunno, i was also wondering about that
[01:29] <Kamion> lamont: is somebody sorting out a live CD release announcement? question on ubuntu-devel@ ...
[01:30] <T-Bone> lamont: ping?
[01:38] <Keybuk> * Considering evolution
[01:38] <Keybuk> ! 3 patch hunks dropped
[01:38] <Keybuk> * Building evolution_2.0.2-3ubuntu1.dsc
[01:39] <Keybuk> that's quite impressive, seeing as it's doing 1.4 -> 2.0 on both sides
[01:44] <lamont> Kamion: mako said he would put something together yesterday...
[01:45] <T-Bone> lamont: i'm facing a lot of "unmet dependencies" errors on stage2...
[01:45] <mako> Kamion: ergh, yes.. got sucked into cc stuff.. will do
[01:48] <Kamion> ta
[01:51] <lamont> Keybuk: impressive, or scary? :-)
[01:51] <lamont> T-Bone: so, um, build them.. :-)
[01:52] <lamont> T-Bone: example missing build-dep?
[01:52] <T-Bone> lamont: heh, i'm scared that if they are missing, it's because they didn't build in the first place ;)
[01:52] <lamont> probably
[01:52] <T-Bone> lamont: i have also stuff much more scary:
[01:52] <T-Bone> cpp-3.3: installed (negative dependency)(but version ok 1:3.3.4-9ubuntu5 << 1:3.3.3-0pre1)
[01:53] <T-Bone> look at "<<"
[01:54] <T-Bone> that's the build log for xfree86. You can take a look at it, it's rather small (3k)
[01:54] <lamont> T-Bone: url?
[01:54] <lamont> or hostname?
[01:55] <Keybuk> lamont: well, it means either my hunk-at-a-time patcher works ... or there's a major bug in it noticing patch failures :p
[01:55] <lamont> Keybuk: yeah.  cool.
[01:55] <T-Bone> lamont: http://envy.esiee.fr/~varenet/logs/xfree86_4.3.0.dfsg.1-6ubuntu25_20041027-1248
[01:56] <lamont> t-bone: the neg dep is fine
[01:56] <lamont> it's the lack of libglide3-dev that's fatal.
[01:56] <T-Bone> ok
[01:56] <T-Bone> yep
[01:57] <T-Bone> but isn't it scary that it says "1:3.3.4 << 1:3.3.3" ?
[01:57] <lamont> it's a build-conflicts
[01:57] <Keybuk> lamont: what's cool is that it spins rookery's pid counter about 6 times on a run <g>
[01:57] <lamont> it could be that dpkg could use a better message, but keybuk is working on the merge, so lets leave him alone right now. :-)
[01:57] <lamont> Keybuk: lol
[01:58] <lamont> T-Bone: how did the stage1 glide build go?
[01:58] <Keybuk> that's an sbuild message, isn't it?
[01:58] <thom> Keybuk: you're a bad man
[01:58] <Keybuk> thom: I am?
[01:59] <T-Bone> lamont: do you know the package name? I can't find a stage1 log for "*glide*" :-/
[01:59] <lamont> t-bone: apt-cache show libglide3-dev | grep ource
[01:59] <lamont> Source: glide
[01:59] <T-Bone> lamont: then i assume it didn't go
[01:59] <T-Bone> [varenet@envy ~] $ grep glide build_list
[01:59] <T-Bone> [varenet@envy ~] $ 
[02:00] <T-Bone> looks like it wasn't kept by quinn-diff
[02:00] <lamont> try explicitly building it against stage1?
[02:01] <T-Bone> will do
[02:03] <pitti> lamont: Hi, howdy!
[02:04] <pitti> lamont: Has the glibc issue been resolved?
[02:04] <fabbione> hey T-Bone 
[02:04] <lamont> pitti: yeah.
[02:05] <pitti> lamont: mdz is sleeping, can you unleash the package to the security archive, too? (and call amber)
[02:05] <lamont> pitti: in debian/control.in/libc6, g/Architecture:/s/:/: amd64/
[02:06] <T-Bone> fabbione: howdy!
[02:06] <pitti> lamont: now it builds also on amd64? Must I upload a package with this change or did you already do this?
[02:06] <lamont> ubuntu2.2 is in the archive
[02:06] <lamont> well, source and binary are sitting there waiting to enter the archive when amber gets goosed
[02:07] <Kamion> U := $(CURDIR)/debian/udev-udeb
[02:08] <Kamion> I just KNOW this is going to confuse me
[02:08] <pitti> lamont: argh, I have to wait for mdz anyway because only he can actually approve the mail to u-s-a
[02:08] <lamont> hoary.all.amd64:Total 1130 package(s) in state Needs-Build.
[02:08] <lamont> hoary.all.i386:Total 472 package(s) in state Needs-Build.
[02:08] <lamont> hoary.all.powerpc:Total 674 package(s) in state Needs-Build.
[02:08] <Kamion> what's stuck with amd64?
[02:09] <lamont> Kamion: me.  only 2 buildds atm, while the other 2 both have 3
[02:09] <Kamion> ah
[02:09] <lamont> that and I had one of the 2 offline for a bit over an hour last night
[02:11] <lamont> Kamion: I need to resurrect king now that it's happy again.
[02:12] <fabbione> T-Bone: how is going the ia64 port?
[02:12] <T-Bone> fabbione: it's going ;^)
[02:13] <T-Bone> i'm experiencing build failures that have to be solved on a case-by-case basis, so that takes time
[02:13] <T-Bone> fabbione: i'm currently progressing in "stage2", aka "warty built against warty"
[02:14] <Kamion> hm ... I bet d-i doesn't have /bin/mountpoint
[02:14] <fabbione> T-Bone: cool. please let me know when you have a working pure warty chroot :-)
[02:15] <T-Bone> fabbione: heh. Don't worry, when that happens, i'll let people know!
[02:15] <fabbione> ehehe
[02:45] <thom> lamont: what's the deal with gnutls?
[02:45] <lamont> thom: ftbfs?
[02:45] <lamont> doesn't seem to be...
[02:46] <thom> lamont: no, no. gnutls11 v gnutls10
[02:46] <sivang> morning folks
[02:46] <thom> are we standardising on 11 for hoary?
[02:47] <lamont> oh that... /me needs to request a sync of gnutls11 from debian (new package and all that...)
[02:52] <elmo> thom: yes
[02:52] <elmo> lamont: no, you don't
[02:52] <elmo> it's there
[02:52] <lamont> thom: metric boatload of changes if we don't, and no reason not to...
[02:52] <lamont> elmo: even better.
[02:52] <elmo> people just need to start merging packages that are being held back
[02:52] <robtaylor> amu: alive yet?
[02:52] <elmo> like libldap2
[02:52] <lamont> DOH!
[02:57] <robtaylor> lamont: hal doesnt seem to work on the livecd...
[02:57] <robtaylor> :(
[02:59] <fabbione> daniels: please nuke Xprint out of X.org
[02:59] <fabbione> daniels: even upstream
[02:59] <fabbione> they don't deserve to live
[03:00] <fabbione> THEY MUST DIE NOW
[03:02] <Micksa> geez
[03:02] <Micksa> you don't like them?
[03:03] <thom> xprint is the stupidest idea in the world. EVAH.
[03:04] <Micksa> what's wrong with the idea?
[03:06] <daniels> fabbione: i'm working on killing it upstream, so is everyone else
[03:06] <fabbione> daniels: thanks
[03:06] <daniels> fabbione: remember how I asked you not to ship it? :)
[03:07] <fabbione> daniels: THEY SHOULD BE BANNED FROM EARTH just for the fact that they use their own site.def and another set of Imake Defines.
[03:07] <lamont> robtaylor: wonder if it's just that hal doesn't get started...
[03:08] <daniels> fabbione: xprint is total bong
[03:08] <lamont> try sudo /etc/init.d/hal start (or whatever it is...)
[03:08] <lamont> enrico: does that mean they'll be the first to get to mars?
[03:09] <enrico> lamont: probably.  So they are the one that have to build the toilets
[03:09] <robtaylor> lamont: nope worse than that... see http://bugzilla.ubuntu.com/show_bug.cgi?id=2758
[03:09] <robtaylor> lamont: basically there seems to be a fundamental disconnect somwhere between the kernel and hald
[03:10] <lamont> robtaylor: like that livecd uses a completely different hw detection mechanism?  would that explain it?
[03:10] <fabbione> lamont: no... they won't be the first anyway
[03:10] <fabbione> lamont: http://www.fabbione.net/bla/marte.jpg
[03:10] <fabbione> marte = mars
[03:11] <lamont> the mars expedition folks here in the states are headquartered in Boulder, CO.
[03:11] <lamont> fabbione: cool pic
[03:11] <Mithrandir> fabbione: sheez, put ut some decent pizza place at least. ;)
[03:11] <robtaylor> lamont: i dont know.. bascially what happens is if you plug in a usb device , hal doesnt even get told of its existance - might be a dbus problem? if the devcie is plugged in at startup, then hal knows about it just fine...
[03:11] <fabbione> ehhehe
[03:12] <lamont> robtaylor: could be.  is dbus et al even running on a booted livecd?
[03:12] <robtaylor> lamont: you mentioned that mini_fo has problems with fs mapped usix sockets? maybe its this?
[03:12] <lamont> tcp sockets
[03:12] <lamont> specifically, polling on same with select
[03:13] <lamont> ==> apt http repositories fail
[03:13] <robtaylor> lamont: yep, dbus is running
[03:13] <robtaylor> lamont: oh, thats very odd
[03:14] <Kamion> oh, shit, I need to be gone. Back in a few hours.
[03:14] <lamont> robtaylor: because dereferencing a null pointer in the fs layer before you _get_ to berkeley sockets is still fatal
[03:14] <robtaylor> Kamion: laters
[03:15] <robtaylor> lamont: sounds pretty scary, one way or another =)
[03:15] <lamont> yeah - still working out the right anagram expansion for minifo
[03:15] <lamont> (NFS == Not a File System"
[03:16] <robtaylor> Might interact nastily in filesystem operations?
[03:16] <lamont> s/interact/interfere/
[03:16] <robtaylor> yeah
[03:16] <robtaylor> =D
[03:17] <robtaylor> i think we got it ;)
[03:17] <lamont> heh.  that's the clean version
[03:17] <justdave> enrico: pong
[03:19] <enrico> justdave: hey! :)
[03:19] <enrico> justdave: solved, thanks
[03:20] <robtaylor> lamont: it just seems that the kernel isnt running hotplug ..
[03:21] <robtaylor> lamont: if i make hotplug.fuctions write stuff to a file whenever its invoked, nothing ever gets written to that file..
[03:22] <robtaylor> unless somewhow the kernel is invoking hotplug in a different root ... is that even possible?
[03:30] <lamont> very
[03:30] <lamont> there's the real root, the basemod root, and the mainmod root.
[03:30] <lamont> pick one. :-(
[03:31] <robtaylor> but surely forat a given point the kernel should be invoking whatevers in /proc/sys/kernel/hotplug in whatever is currently mounted as root?
[03:31] <robtaylor> s/forat/at
[03:32] <lamont> hence mdz's comment about chroot vs pivotroot
[03:32] <robtaylor> hmm. anyway i just tried touching /proc/sys/kernel/hotplug to see if it was a cached inode problem, but that didnt seem to fix it?
[03:32] <robtaylor> s/?/.
[03:33] <robtaylor> ok, laters lamont
[03:33] <lamont> this up at 0530 crap is really annoying this night person. :-)
[03:33] <lamont> :-(
[03:33] <lamont> back in an hour or two
[03:33] <robtaylor> heh. slerep tight
[03:34] <fabbione> lamont: are you trying to turn into me, waking up at 5am each day?
[03:35] <lamont> fabbione: alarm clock goes off at 0530 every school day.  kids leave for school 0645.  If I drive, then I'm usually back at 0800, if I don't (like today), then I sometimes fall back into bed.
[03:36] <fabbione> oh yeah...
[03:36] <fabbione> forgot about the kids going to school
[03:36] <lamont> otoh, my split up day does make it easier for me to catch the non-US crowd(s).
[03:37] <lamont> anyway,  bbl
[03:37] <fabbione> later
[04:27] <thom> *sigh*; firefox has 24 bugs
[04:29] <Mithrandir> only 24?
[04:29] <thom> Mithrandir: most of them are upstream or weird crashers with flash or java
[04:31] <Mithrandir> probably "if you feed it this invalid bytecode, the JVM will do funny things and crash FF"
[04:58] <fabbione> AHHHHHHHH
[04:58] <fabbione> IT DOESN'T EVEN RESPECT THE OPTION TO NOT BUILD ITSELF!
[05:05] <fabbione> Mithrandir: thanks :-)
[05:05] <fabbione> root access at fd.org is enough
[05:05] <Mithrandir> we could just sit on daniels next time we meet him, until he gives us?
[05:05] <fabbione> rm -rf /var/lib/cvs/xorg/xc/programs/Xserver/XpConfig would be enough
[05:06] <fabbione> Mithrandir: eheheh
[05:13] <thom> ok. firefox in warty is 0.99+1.0PR.1+revertedto0.9.3-0ubuntu3
[05:13] <elmo> seb128: ?
[05:13] <seb128> elmo: ?
[05:13] <thom> is 1.0~0.10.1-0ubuntu1 a reasonable version for hoary? (can we use ~ yet?)
[05:14] <elmo> seb128: you're only doing ubuntu uploads were the are ubuntu changes, right ?
[05:14] <mdz> gah, did pitti leave?
[05:14] <elmo> thom: I don't know if we want to, but if the consensus is that you do, let me know, because I'll have to unblock it in katie
[05:15] <seb128> elmo: hum ? I've uploaded new versions that are not and debian, and merged warty changes for the *gksu*
[05:15] <daniels> thom: hmm, since really < reverted, you could have 0.99+1.0PR.1+seriously+1.0PR.1-0ubuntu1
[05:15] <elmo> seb128: ok, I couldn't tell from your changelog, if there still warty changes or not.. just wanted to double check
[05:15] <seb128> elmo: ok, that's fine, thanks for checking :)
[05:16] <daniels> or 'totally+1.0PR.1'
[05:17] <thom> i really don't want to go down the 0.999 road
[05:18] <sid77> 0.999 is evil (almost)
[05:19] <daniels> sid77: as opposed to 0.99+1.0PR.1+seriously+1.0PR.1?
[05:19] <aj> elmo: does dak logging/etc deal with ~ okay now?
[05:20] <sid77> daniels, that is evil (completely)
[05:20] <thom> so the quick answer was "no, you can't use ~. KTHXBYE"
[05:20] <thom> ;-)
[05:20] <elmo> aj: no, that was implicit in "unblock", i.e. use something other than tilde-seperated-value for logging :)
[05:21] <aj> elmo: heh
[05:21] <elmo> thom: no, the katie thing can and should be fixed - the answer to your question is independent to that
[05:22] <aj> thom: "1.0~0.10.1-0ubuntu1" is hella complicated; the 0.10.1 can't be just pre1?
[05:24] <thom> aj: probably, yes
[05:26] <thom> mdz: ^ ?
[05:27] <mdz> thom: ?
[05:27] <thom> mdz: thoughts on firefox versioning?
[05:27] <mdz> thom: for going back to 1.0PR?
[05:27] <thom> yeah
[05:27] <Keybuk> 1.0~PR ?
[05:27] <thom> (with appropriate patches)
[05:28] <thom> Keybuk: my suggestion was close to that, yes
[05:28] <jdub> tilde?
[05:28] <Keybuk> jdub: sorts less than null, so 1.0~ < 1.0
[05:28] <Mithrandir> jdub: tilde mean "less than" in newage debian versioning speak.
[05:28] <jdub> heh
[05:29] <Mithrandir> or less than zero
[05:29] <Mithrandir> minus epsilon
[05:29] <Mithrandir> or something
[05:29] <Keybuk> it's a bugger to explain, isn't it <g>
[05:29] <aj> "actually, immediately prior to what i just said" kinda works
[05:30] <aj> "it's version 1.0 -- actually just before 1.0 -- sub version pre1"
[05:30] <jdub> perhaps "... but i'm bullshitting you" works
[05:31] <aj> nah, better to reserve that for the 0.999-but-really-1.0pre1 versions out there
[05:32] <mdz> thom: '~' gives me the fright
[05:33] <Keybuk> mdz: any particular reason?
[05:33] <aj> mdz: ooo, costume ideas for halloween
[05:33] <mdz> Keybuk: because nobody has used it for anything particularly important yet
[05:33] <Keybuk> that's because the Debian archive won't accept them until dpkg 1.10 is in a stable release
[05:33] <mdz> but of course, now is the time to break it if it 's going to break...
[05:33] <mdz> it needs an elmo signoff
[05:34] <Keybuk> I was kinda nervous, then I read the code and it was a "woah! that's actually really elegant" when I saw how it was implemented
[05:34] <aj> /msg elmo quick: type /quit ~ is okay by me!
[05:34] <aj> Keybuk: (of /course/ it was really elegant...)
[05:35] <fabbione> hey mdz
[05:35] <aj> (i mean, really -- code getting into dpkg without being elegant? the very thought!)
[05:35] <Keybuk> aj: yeah, I mean, it's so great that I could merge the bzip2 stuff in and being totally sure it would just work
[05:36] <Keybuk> the fact it didn't even *compile* because someone typo'd the #ifdef so never actually ever ran it was a figment of my imagination
[05:36] <aj> Keybuk: i guess you take drugs for that now though?
[05:37] <Keybuk> aj: I must be, for agreeing to maintain the fucking thing :p
[05:37] <Mithrandir> Keybuk: did you ever agree to it?
[05:37] <aj> Keybuk: oh dear, never a good sign when you don't know what drugs you're taking
[05:37] <Mithrandir> I thought we just made you?
[05:38] <elmo> mdz: I'll need to unbreak katie - but I thought you had concerns about back compat or so?
[05:39] <mdz> elmo: depends on whether sarge releases before hoary
[05:39] <mdz> well
[05:39] <mdz> mozilla-firefox isn't even in woody, is it?
[05:39] <Keybuk> no
[05:39] <Keybuk> it didn't *exist* when woody was released
[05:40] <jdub> (haw haw)
[05:40] <aj> i don't see how it matters even if hoary releases before sarge?
[05:40] <mdz> aj: if hoary releases before sarge, we will support upgrades from woody to hoary
[05:40] <aj> you've got 1.0.1 in woody, 1.1 in warty, 1.2~1 in hoary, 1.2 in sarge
[05:41] <aj> your only bug is if someone uses woody's dpkg, and installs the package from hoary or sarge, then tries pointing apt and the other, without ever upgrading dpkg/apt
[05:41] <Keybuk> mdz: but we don't support upgrades from woody to warty ?
[05:43] <mdz> Keybuk: of course we do
[05:44] <Micksa> lifeless: I'm still undecided on whether to come on on friday.
[05:44] <Micksa> oops.
[05:44] <Micksa> grah
[05:45] <mdz> thom: why don't we stick with what Debian is doing?  they seem to have 0.10.1+1.0PR
[05:45] <mdz> that's newer than 0.99+1.0PR.1+revertedto0.9.3-0ubuntu3
[05:45] <Mithrandir> 0.100.1+1.0PR ?
[05:45] <Keybuk> it really isn't dude
[05:45] <Keybuk> 0.10 < 0.99
[05:46] <mdz> oh, missed the 0.99
[05:46] <thom> 0.10.1 so isn't hiugher
[05:46] <mdz> wtf is 0.99?
[05:46] <mdz> I read it as 0.9
[05:46] <Keybuk> mdz: nearly but not quite
[05:46] <sid77> what about a 0.9.3-<current version>?
[05:46] <Keybuk> sid77: 9 < 99
[05:46] <thom> 0.99 was my screwup early on
[05:46] <mdz> thom: if elmo is ok with ~, I'm ok with ~
[05:47] <mdz> elmo: you need to unbreak katie, or you need to remove the fascist check that forbids ~?
[05:47] <mdz> I thought katie used apt's version comparison code
[05:47] <sid77> Keybuk, right, but doesn't it start from 0.9.3?
[05:47] <elmo> mdz: it does - I thought apt supported ~ ?
[05:47] <mdz> elmo: it does
[05:48] <mdz> elmo: so what unbreaking is required?
[05:48] <thom> mdz: 16:20 < elmo> aj: no, that was implicit in "unblock", i.e. use something other than tilde-seperated-value for logging :)
[05:48] <Keybuk> sid77: 0.9 is less than 0.99 because 9 is less than 99
[05:48] <elmo> mdz: I use ~ as a separator in log files.   go me.
[05:48] <mdz> elmo: oh, rad
[05:48] <elmo> and also there's a regex which checks for "validity" which needs updated, but that's trivial
[05:48] <thom> yay for intra debian communication
[05:48] <thom> ;-)
[05:49] <mdz> "implement"?
[05:49] <elmo> thom: it's nothing to do with intra debian communication, tilde-seperated log files are something I picked up from work, and at the time I wrote the katie stuff, ~ wasn't even being discussed for versions
[05:49] <elmo> err, previous work, obviously
[05:49] <Keybuk> we're fast running out of silly characters to use for things
[05:49] <sparkes> yay, utf
[05:50] <mdz> \000
[05:50] <sid77> Keybuk, maybe we should introduce a new number space where 99 < 9 :)
[05:53] <Keybuk> thom: does firefox use libtool anywhere?
[05:53] <lifeless> sid77: aleph-1 ?
[05:53] <lifeless> (as opposed to aleph1)
[05:54] <thom> Keybuk: oh christ
[05:54] <sid77> lifeless, geeee...
[05:54] <mdz> elmo: is it realistic to change that in order to allow for thom's firefox upload? or do we need to do something else for the short term?
[05:55] <elmo> mdz: I can probably fix it tonight - shouldn't take long
[05:55] <mdz> Keybuk: will libtool require un-breaking as well?
[05:55] <Keybuk> yes
[05:56] <mdz> hell
[05:56] <mdz> Keybuk: how bad is it?
[05:56] <elmo> why would libtool care about the Debian version ?
[05:56] <Keybuk> mdz: you know Libtool just flatly falls on its face and flops about when you have spaces in file or directory names?
[05:56] <thom> elmo: the ~ is in the upstream version?
[05:57] <Keybuk> elmo: mozilla-firefox-1.0~1/debian/...
[05:57] <elmo> this is going in the filename??
[05:57] <Keybuk> elmo: goes in the directory name, doesn't it
[05:57] <mdz> Keybuk: sure, but we're not giving it spacess, we're giving it tildes :-)
[05:57] <thom> ok, so maybe i'll work on some other stuff for a while :-)
[05:58] <Keybuk> mdz: libtool uses IFS="~" internally for some things
[05:58] <mdz> err
[05:58] <Keybuk> (multi-command link lines)
[05:58] <mdz> it's starting to sound like it would be easier to fix dpkg and apt :-P
[05:58] <seb128> jdub: 
 cneumair * gnome-panel/gnome-panel/ (10 files): (log message trimmed)
 Fix bug #143963 [http://bugzilla.gnome.org/show_bug.cgi?id=143963] : "Remove Distro Menu".
[05:58] <seb128> jdub: finally :p
[05:58] <jdub> hooray! :)
[06:06] <Keybuk> thom: http://people.ubuntu.com/~scott/keybuk-tilde-beep.patch is the libtool patch
[06:07] <daniels> Keybuk: dude, BEL?  seek professional help.  like, three weeks ago.
[06:07] <Keybuk> daniels: well, it's unlikely to appear in filenames <g>
[06:08] <Keybuk> (well, unless your cat walks across your keyboard while in vi)
[06:08] <daniels> libtool has destroyed your brain
[06:09] <thom> holy mother of god save us.
[06:09] <Keybuk> it was ^M, but some editors "helpfully" strip that
[06:10] <aj> Keybuk: it'd be possible to have "1:1.0~pre1" be translated to "1.0pre1" for user visible stuff, like filenames, directory names, dselect, etc. that'd make things like 1.0~1 suck, of course.
[06:10] <Keybuk> aj: that's the other solution, yes.  fix dpkg-source
[06:11] <Keybuk> that's actually less sick than it sounds, because we already don't put the epoch and ':' in filenames
[06:11] <aj> apt does, but url escapes them
[06:11] <Keybuk> tar gets somewhat upset when asked to do 'tar czf blah_1:1.0.orig.tar.gz blah-1:1.0'
[06:11] <aj> really?
[06:11] <Keybuk> sure
[06:12] <aj> why?
[06:12] <Keybuk> it thinks blah-1 is a hostname to rcp 1.0 from
[06:12] <aj> really??
[06:12] <Keybuk> yup
[06:12] <Keybuk> you've never encountered this before?! :p
[06:12] <aj> i don't use colons in filenames :)
[06:12] <aj> nor do i use rsh :)
[06:12] <Keybuk> this is the reason dpkg hides the epoch from the filenames
[06:12] <Keybuk> it breaks things :p
[06:14] <thom> that sounds like a much saner approach for ~ too
[06:14] <aj> wow, tar sucks
[06:14] <elmo> Keybuk: WTF dude, neither had you till a week ago - I told you about it in London :-P
[06:14] <Keybuk> elmo: true :o)
[06:14] <aj> geez, elmo's such a spoil sport
[06:14] <Keybuk> that was the first time I'd encountered it <g>
[06:14] <Keybuk> and, admittedly, probably the first time I'd ever tried to use ':' in a filename
[06:16] <aj> tar czvf - foo:bar > foo:bar.tgz works though :-/
[06:16] <aj> ahh, gnu tools. gotta love 'em.
[06:19] <robtaylor> Keybuk: s/~/LESSTHANONOTREALLYMAYBE/ ;~)
[06:19] <robtaylor> lets keep that clear semantic information intact ;-) 
[06:22] <jdub> night all
[06:24] <robtaylor> night jdub 
[06:25] <Keybuk> %7F is ugly
[06:25] <Keybuk> dpkg-source.pl: extracting build-essential in build-essential-11%7F
[06:26] <Keybuk> much cuter without it
[06:26] <thom> yes
[06:26] <thom> that is quite nasty
[06:27] <Keybuk> dpkg-source.pl: extracting build-essential in build-essential-11_
[06:27] <Keybuk> a little nicer
[06:27] <Keybuk> bleh
[06:28] <robtaylor> heh. more mathematicallys s/~/)/
[06:28] <robtaylor> ;)
[06:30] <robtaylor> (out of interest.. whats the problem with just using ~ in the filename?)
[06:32] <Keybuk> robtaylor: libtool go boom
[06:32] <robtaylor> Keybuk: ah . boh
[06:34] <Keybuk> though I actually can't make it fail now, amusingly
[06:34] <robtaylor> Keybuk: is there a particular reason it might choke on ~?
[06:34] <Keybuk> robtaylor: yeah, uses IFS=~ internally
[06:35] <robtaylor> ahhh
[06:35] <robtaylor> ugh
[06:35] <Keybuk> though it's now working on places it should fail
[06:36] <Keybuk> ar cru .libs/libtest.a  /home/scott/tmp/tilde~test/test2.o
[06:36] <Keybuk> ranlib .libs/libtest.a
[06:36] <Keybuk> -- 
[06:36] <Keybuk> that should fail, theoretically
[06:38] <Keybuk> maybe it's just 1.4 which breaks, in which case I don't care
[06:38] <Keybuk> (you can see I really tested this, can't you :p)
[06:38] <robtaylor> heheh
[06:38] <robtaylor> isnt that waht users are for? ;)
[06:40] <Keybuk> thom: can you try building firefox in a dir with ~ in it ... I have a hunch it'll work
[06:41] <Keybuk> my libtool test suite *cough*gnome*cough* builds ok
[06:41] <thom> heh
[06:41] <Keybuk> 1.4 does fail
[06:42] <Keybuk> but 1.5 actually doesn't eval the variables until inside the loop (at which point the IFS=~ doesn't matter)
[06:43] <Keybuk>     save_ifs="$IFS"; IFS='~'
[06:43] <Keybuk>     for cmd in $cmds; do
[06:43] <Keybuk>       IFS="$save_ifs"
[06:43] <Keybuk>       eval cmd=\"$cmd\"
[06:44] <Keybuk> so the variable with the directory name in it doesn't get expanded inside $cmd until afterwards
[06:44] <Keybuk> cool
[06:44] <bluefoxicy> o.o
[06:45] <bluefoxicy> Where would i go to discuss proactive security (http://wiki.ubuntu.com/ProactiveSecurity) and the development of Ubuntu (and Debian by upwards inheritance) to include enhancements not detrimental to the user experience or to binary compatibility?
[06:56] <Keybuk> gah, pretty initscripts make merging hurt
[06:59] <Keybuk> zsh: segmentation fault (core dumped)  ls -bFh --color=auto -d ../sources/*/results
[06:59] <Keybuk> gee... thanks
[07:03] <thom> Keybuk: it's still running, but it's not blown up
[07:04] <Keybuk> elmo: do you want a list of things we can actually just sync from Debian again?
[07:05] <Keybuk> most of lamont's libtiff4-fest for example
[07:06] <Kamion> hm, I do like the way my random test Ubuntu installs are set up with a random language
[07:06] <Kamion> and keyboard layout; makes for extra fun when typing passwords
[07:06] <Keybuk> heh
[07:07] <thom> heh
[07:12] <Kamion> hooray, my udev-ised test CD ... totally blows up
[07:13] <thom> rock on
[07:14] <mdz> Keybuk: do you have a merged alsa-utils?
[07:15] <Keybuk> yup, want it?
[07:15] <mdz> yup
[07:15] <thom>  < Tybstar> it would be cool if CUPS had a reasonable default paper size depending on the location you picked at install
[07:16] <mdz> thom: yes, it would
[07:16] <thom> (he works for specifix, who do conary)
[07:17] <thom> oh dear:
[07:17] <thom> 18:15 < sri> fer: just think how awsome the porn would be, various actions related to each stage woo..
[07:17] <thom> 18:15 < fer> sri: wops
[07:17] <thom> (on graphical boot)
[07:18] <thom> 18:16  * sri can imagine hww delicious fscking would be :)
[07:18] <thom> 18:16 < fer> linux-2.6.9-p0rn1
[07:18] <thom> 18:16 < fer> and mount!
[07:18] <Keybuk> mdz: http://people.ubuntu.com/~scott/alsa-utils/
[07:19] <Keybuk> there's an oddness there, it looks like Debian dropped 90_debian_alsaconf.dpatch
[07:19] <Kamion> ah, busybox mount doesn't support --bind, hmm
[07:19] <Keybuk> and there's a few Debian changes to control in .dropped that didn't apply
[07:23] <mdz> thom: CRACK
[07:23] <thom> mdz: *g*
[07:23] <mdz> Keybuk: if debian dropped that patch, why is it still in debian/patches in the output?
[07:23] <Keybuk> mdz: because we changed it
[07:23] <mdz> we what?
[07:23] <Keybuk> we changed 90_debian_alsaconf.patch
[07:24] <Keybuk> -+  cfgfile="/etc/alsa/modutils/1.0"
[07:24] <Keybuk> ++  cfgfile="/etc/modutils/alsa-base"
[07:24] <Keybuk> specifically
[07:24] <mdz> ok, that's not important, since we later dropped alsaconf entirely
[07:24] <Keybuk> bit trivial really :)  should be easy to resolve, find which of the tree split files that's in, fix it, and kill the warty file 
[07:24] <mdz> I wonder if it's fixed in 1.0.6
[07:25] <Keybuk> _warty.patch is the base->warty patch, _debian.patch is base->debian, _hoary.patch is debian->merge
[07:26] <Keybuk> so you pretty much side-by-side compare _warty.patch and _hoary.patch to see whether anything went wrong
[07:27] <mdz> where did all these control file changes come from?
[07:27] <mdz> they're not in the changelog
[07:27] <Keybuk> "fabbione" :)
[07:27] <Keybuk> the ones in .dropped are *Debian* changes
[07:27] <Kamion> mmmkay, I wonder if d-i trunk is hosed at the moment
[07:28] <Keybuk> the ones in _warty.patch were our changes
[07:28] <mdz> where is .dropped?
[07:28] <Kamion> there's some very suspicious pissing about with /proc/self/fd/0 and /dev/console
[07:28] <Keybuk> same directory
[07:28] <Keybuk> http://people.ubuntu.com/~scott/alsa-utils/alsa-utils-1.0.6.dropped
[07:28] <mdz> ah, my 'get *' didn't get it
[07:28] <mdz> oh, yes it did
[07:28] <Keybuk> that basically logs the IKEA "I put it together, what's this bit for?" hunks
[07:28] <mdz> I thought you meant literally ".dropped"
[07:29] <Kamion> #6
[07:29] <Kamion> (oops)
[07:29] <bluefoxicy> o_o
[07:30] <bluefoxicy> hmm.
[07:30] <bluefoxicy> Noo ne?
[07:30] <mdz> Keybuk: the reason that one fell out is because fabio changed the Maintainer: field in the ubuntu branch for some reason
[07:30] <Keybuk> yup
[07:31] <Keybuk> not sure why the description fell out, I suspect just because it was a bit too close
[07:31] <thom> bluefoxicy: the ubuntu devel mailing list would seem to be appropriate
[07:34] <mdz> Keybuk: resolved, uploaded
[07:34] <Keybuk> cool, I'm just finishing off the last of the packages without a .dropped (basically checking things are sane)
[07:34] <mdz> Keybuk: did you get official word from elmo about your key?
[07:34] <Keybuk> then I'll stick them all somewhere
[07:34] <Keybuk> mdz: not yet
[07:35] <mdz> mdz@jackass:~ $ gpg --keyring /srv/keyring.no-name-yet.com/keyrings/ubuntu-keyring.gpg --list-keys remnant
[07:35] <mdz> pub  1024D/84AD676C 1999-12-30 Scott James Remnant <scott@netsplit.com>
[07:36] <mdz> looks to me like it's there, as it should be
[07:36] <Keybuk> that's the wrong one
[07:36] <bluefoxicy> thom:  oy now I have to find those-
[07:36] <Keybuk> C978C8AE  is my canonical key
[07:36] <bluefoxicy> thom:  no better suggestions?
[07:40] <mdz> Keybuk: does there exist a copy of that key with useful signatures on it?
[07:41] <mdz> I seem to have one in my keyring, but it has zero signatures
[07:41] <seb128> hum. Any way to get a sync for librsvg2 2.8.1-1 ? The warty changes are in the new debian version too 
[07:41] <mdz> seb128: sure, just email elmo
[07:42] <Keybuk> mdz: should be signed by daniels, fabbione, lalo and debonzi
[07:42] <mdz> Keybuk: but where?
[07:42] <mdz> it's not on pgp.net or keyring.d.o
[07:42] <mdz> Keybuk: if you only keep the signatures locally, it doesn't count :-)
[07:44] <Keybuk> subkeys.pgp.net has them
[07:47] <seb128> elmo: any way to get a sync from debian for librsvg2 2.8.1-1 ? The warty changes are in the new debian version
[07:48] <elmo> seb128: done
[07:48] <seb128> thanks !
[07:50] <bob2>  hm, why is "apache-common" in main, while the rest of apache 1 is in universe?
[07:51] <elmo> it's not
[07:51] <elmo> apache-dev is in main too
[07:51] <elmo> for php4
[07:51] <elmo> (b-d)
[07:51] <bob2> hrm, ok
[07:51] <bob2> well, the apache 1 daemon is in universe at least
[07:52] <bob2> oh, ew, some of apache 1 is in main so libapache2-mod-php4 can build?
[07:53] <Kamion> hm, current grub seems to build-dep on type-handling too ...
[07:54] <mdz> that type-handing crack needs to go
[07:55] <mdz> grub.maintainer == type-handling.maintainer, I think
[08:02] <Kamion> mdz: yes
[08:02] <Kamion> mdz: you misspelt "dealer"
[08:03] <mdz> lamont: ping?
[08:03] <mdz> Kamion: are you running britney or anything like that over hoary yet?
[08:03] <mdz> do we have some idea what the dependency breakage is like?
[08:04] <Kamion> mdz: elmo runs it now, I just mirror it
[08:04] <Keybuk> -Maintainer: Tommi Virtanen <tv@debian.org>
[08:04] <Keybuk> +Maintainer: Fabio M. Di Nitto <fabbione@fabbione.net>
[08:04] <Keybuk> heh
[08:04] <elmo> mdz: yeah, it's bad, 'cos of gnutls mostly AFAICS from a quick glance
[08:04] <elmo> which should be building now
[08:04] <__daniel> britney?
[08:04] <Kamion> mdz: but it's in the same place, yes
[08:04] <Kamion> http://people.ubuntulinux.org/~cjwatson/testing/
[08:04] <mdz> Keybuk: which package is that?
[08:05] <mdz> Kamion: found it, thanks
[08:05] <Kamion> __daniel: script that manages Debian's testing distribution; we run it in a very cut-down form to check for dependency breakage
[08:05] <mdz> elmo: gnutls caused binary deps to break?  or build-deps?
[08:05] <elmo> mdz: binary, gnutls10 -> 11 transition
[08:06] <Keybuk> mdz: uh, python-osd I think
[08:06] <elmo> e.g. libldap2 is uninstallable
[08:06] <Keybuk> I went past it pretty quickly with a giggle
[08:06] <__daniel> Kamion: wow - that sounds cool - every time i hear about debian / ubuntu infrastructure i'm absolutely impressed
[08:06] <mdz> ah, right, all the libgnutls10 deps on our existing binaries
[08:06] <mdz> that sort of sucks
[08:07] <Kamion> __daniel: was kind of a necessity when we were putting things together, otherwise the only way to tell whether warty was hosed was to build a CD, rsync it down, burn it, boot/install it, and see if it worked
[08:08] <Kamion> __daniel: I got bored of doing that very quickly :)
[08:09] <__daniel> kamion: that's what mvo_ told me (he lives just some km s away from me): "after some time, you just see, what the infrastructure DOESN'T do for you"
[08:09] <__daniel> i'm still very perplexed by it all :-)
[08:28] <mdz> elmo: so once gnutls is built, we need to rebuild pretty much everything?
[08:38] <lamont> mdz: ack
[08:39] <mdz> lamont: what does the hoary buildd picture look like?
[08:40] <lamont> hoary.all.amd64:Total 894 package(s) in state Needs-Build.
[08:40] <lamont> hoary.all.i386:Total 1 package(s) in state Needs-Build.
[08:40] <lamont> hoary.all.powerpc:Total 181 package(s) in state Needs-Build.
[08:40] <lamont> working on that amd64 thing now
[08:41] <lamont> which is part of why it's behind - took one buildd offline to clone the 4 chroots
[08:42] <mdz> lamont: thanks
[08:43] <mdz> lamont: also, how many packages will we need to rebuild to fix the gnutls situation?
[08:44] <lamont> gnutls11 is in the archive, with stuff depwaitng on it - you're after that count?
[08:44] <lamont> hoary.all.amd64 18
[08:44] <lamont> hoary.all.i386 21
[08:44] <lamont> hoary.all.powerpc 20
[08:45] <lamont> packages dep-wait libgnutls11-dev
[08:49] <mdz> lamont: I mean, we have a bunch of pakcages which depend: libgnutls10
[08:49] <mdz> do all of those have new versions which build-dep on the new gnutls?
[08:49] <mdz> or will we need to build some of them explicitly?
[08:50] <lamont> ah, will check.
[08:50] <mdz> thanks
[08:51] <lamont> fwiw, gnutls10 is there, but I gather we want it to go away.
[08:52] <lamont> 70+packages with the depends...  I'm going to let the dust settle on the buildd and then figure out where we are, I think.
[08:52] <lamont> meanwhile, fixing amd64
[08:54] <mdz> lamont: gnutls10 binaries will vanish once gnutls11 binaries appear, no?
[08:55] <mdz> lamont: from the numbers you gave above, it looks like i386 dust is quite settled already
[08:56] <elmo> mdz: no
[08:56] <elmo> it's two separate source packages
[08:56] <mdz> oh
[08:56] <T-Bone> lamont: i'm currently under 50% packages successfully built on stage2
[08:56] <mdz> so we don't actually have a problem, then, do we?
[08:57] <mdz> the stuff which depends on gnutls10 will be satisfied
[08:57] <elmo> mdz: e.g. libldap2 is uninstallable
[08:58] <elmo> because although both are there, they aren't co-installable
[08:58] <elmo> so we need to get our libraries transitioned over ASAP
[08:58] <mdz> oh
[08:58] <mdz> I thought Debian did this transition for sarge
[08:59] <mdz> is it blocked by packages which need merging?
[08:59] <Kamion> there were a few transitions involved, I believe
[08:59] <mdz> ah, openldap2 is branched
[08:59] <elmo> right
[09:00] <elmo> I mentioned this last night - probably should filed a bug, sorry
[09:00] <mdz> Keybuk: openldap2?
[09:00] <elmo> btw - does anyone have any clever ideas on how to fix the problem that we have two suites and the same arch/version of a package wants to be in different components in each suite?
[09:01] <Kamion> anyone know what's responsible for starting udevd?
[09:01] <Kamion> I'm booting a d-i CD here with init=/bin/sh, and udevd is started right at the very beginning of the initrd; I can't figure out what's doing it
[09:02] <lamont> elmo: python? :-)
[09:02] <mdz> openldap2 -2ubuntu1 and -2ubuntu2 seem to be present in Debian
[09:02] <Kamion> component overrides? :-)
[09:02] <lamont> sounds like the component tuple needs to grow by suite-name...
[09:02] <lamont> mdz: do I want to know who the uploader was?
[09:03] <elmo> nono
[09:03] <elmo> remember the pool is split by component
[09:03] <mdz> lamont: no, I mean the changes
[09:03] <mdz> -2ubuntu1 and -2ubuntu2 were ubuutu backports from Debian
[09:03] <Kamion> elmo: oh, toss
[09:03] <elmo> so say we have dpkg_1.9.21 in warty/main, if it's demoted to universe in hoary/, it has to physically move from pool/main, to pool/universe
[09:03] <Kamion> elmo: hardlinks?
[09:03] <elmo> oh, hmm, then again.. why does it? :>
[09:03] <mdz> Kamion: weird, re: udevd
[09:04] <Kamion> elmo: mirrors would probably kind of like it to
[09:04] <mdz> Kamion: I didn't think anything other than S04udev started udevd
[09:04] <Kamion> at least people who don't mirror pool/universe/ (in the vice-versa situation)
[09:04] <lamont> mdz: ah, that helps
[09:04] <Kamion> elmo: such as, hm, just for example, cdimage :-)
[09:04] <Keybuk> mdz: yeah, there's a openldap2 -- am going to re-run these though with the new .po handling Kamion's just helped me with
[09:04] <mdz> Keybuk: openldap2 doesn't have any .po file changes
[09:05] <Keybuk> ok, hang on then
[09:05] <Kamion> mdz: yeah, I'm stumped
[09:06] <Keybuk> http://people.ubuntu.com/~scott/openldap2/
[09:06] <mdz> Kamion: maybe the udev hotplug hook starts udevd if it isn't running?
[09:06] <mdz> Keybuk: fetching, thanks
[09:06] <Keybuk> note that the stuff in .dropped *isn't* identically present in Debian
[09:07] <Keybuk> (though if you're lucky it's a silly whitespace change or something)
[09:09] <Kamion> mdz: the string "udevd" doesn't appear anywhere in the initrd
[09:09] <mdz> Kamion: er, so udevd isn't actually present?
[09:09] <Kamion> /sbin/udevd is there
[09:09] <mdz> that counts as a string :-)
[09:09] <Kamion> no other occurrence of udevd in filenames though
[09:09] <Kamion> er, files
[09:10] <mdz> Kamion: what's in /etc/hotplug.d?
[09:10] <Kamion> chandev default dock ieee1394 input net pci scsi usb
[09:10] <Kamion> ah, default/10-udev.hotplug is a symlink to udevsend
[09:11] <Kamion> that probably explains it
[09:11] <mdz> Kamion: that's what I meant by: <mdz> Kamion: maybe the udev hotplug hook starts udevd if it isn't running?
[09:11] <Kamion> right, I missed it 'cos it was a symlink
[09:12] <mdz>        If udevd isnt already running, udevsend will start it.
[09:12] <mdz> (udevsend(8))
[09:12] <Kamion> is there a way to tell it to queue events instead?
[09:12] <mdz> Kamion: kill -STOP udevd? :-)
[09:12] <Kamion> no, I have to make it go away
[09:12] <mdz> Kamion: you could disable hotplug
[09:12] <Kamion> it's stopping me umounting /initrd
[09:13] <Kamion> from where? :)
[09:13] <mdz> sysctl
[09:13] <Kamion> run from where?
[09:13] <Kamion> this is *before /sbin/init*
[09:13] <mdz> seb128: here?
[09:13] <mdz> Kamion: *blink*
[09:13] <seb128> mdz: yes
[09:14] <Kamion> mdz: or, at least, before I can guaranteeably do anything; umount /initrd is the second command in /sbin/init, and it's failing because udevd is running
[09:14] <mdz> seb128: gnome-settings-daemon segfaults in hoary
[09:14] <mdz> seb128: is this expected?
[09:14] <seb128> mdz: no
[09:14] <seb128> let me upgrade my test box 
[09:15] <Kamion> maybe I could install hotplug somewhere different and move it into place after pivot_root or something
[09:15] <mdz> seb128: /msg'd you the backtrace
[09:15] <Kamion> I can't wait until anna
[09:15] <seb128> mdz: ok, thanks
[09:16] <mdz> Kamion: is there a way to set a sysctl from the kernel command line?
[09:16] <Kamion> mdz: hm, the kernel command line is already running badly out of space for preseeders, don't want to stress it any more
[09:16] <mdz> Kamion: you could just fuser -mk before unmounting it
[09:16] <Kamion> no fuser, and not sure I want to risk killing udevd in the middle of handling an event
[09:17] <mdz> if it's running in something that isn't the real root anymore, should be fairly harmless
[09:17] <Kamion> yeah, but it might mean some bit of hardware doesn't get recognised
[09:17] <Kamion> or?
[09:17] <Kamion> I always thought boot events were queued
[09:19] <mdz> Kamion: later on in the boot process, the udev init script tells it to create nodes for everything it already knows about
[09:20] <Kamion> I think it's probably going to be more predictable to ensure hotplug is not run in the initrd
[09:20] <mdz> I imagine that in a normal system, the events are just ignored in the initrd, since it doesn't have hotplug
[09:20] <Kamion> yeah
[09:26] <seb128> mdz: control-center crash fixed in 2.8.1-0ubuntu2 just uploaded
[09:27] <mdz> thanks
[09:27] <Kamion> mdz: lucky you backed out that perl-in-hotplug thing actually, otherwise trying to use it in d-i would be a total bust
[09:27] <seb128> np
[09:28] <elmo> kamion: ah, yeah, damn, fair point
[09:29] <Kamion> I suspect there'll be problems anyway, /etc/hotplug/*.rc uses a load of stuff that I bet isn't available in busybox
[09:31] <Kamion> elmo: since it needs to be available in both places, hardlinks/copies are probably the only option :(
[09:31] <elmo> yeah, but the whole point of pools was to avoid managing hardlink farms.  gar, sucks.
[09:31] <Kamion> yeah, I know :(
[09:33] <Kamion> mdz: hm, bright spark here has just realised that it might be kind of useful to have modules.pcimap available; hotplug does need that even in 2.6, right?
[09:33] <Kamion> modules.*map I guess
[09:33] <mdz> yes
[09:33] <Kamion> arrrr, this sucks, they're huge
[09:33] <mdz> crap
[09:33] <mdz> I didn't type that fast enough to evade workrave
[09:33] <Kamion> although only 63K compressed, I guess
[09:34] <mdz> I haven't figured out its algorithm yet, but it has a certain small tolerance built into it
[09:35] <Kamion> maybe I can figure out a reduction algorithm; like, only include the lines that actually match modules available in d-i :-)
[09:37] <Kamion> there is a problem with amd64: the boot messages are visible for about a tenth of a second :P
[09:37] <mdz> Kamion: gzipping them wouldn't be a half bad idea, except that hotplug reads them about 80 BILLION TIMES
[09:38] <Kamion> mdz: don't care so much about uncompressed size for now, although the lowmem people will care
[09:38] <mdz> I'm surprised they don't compress better than that
[09:38] <mdz> mdz@potpal:/lib/modules/2.6.8.1-3-386 $ cat *map |wc -c
[09:38] <mdz> 324028
[09:38] <mdz> mdz@potpal:/lib/modules/2.6.8.1-3-386 $ cat *map |gzip -9 |wc -c
[09:38] <mdz> 15500
[09:39] <Kamion> oh, I was using modules.*
[09:39] <mdz> not entirely sure whether modules.alias is used still
[09:39] <mdz> modules.dep you need anyway, of course
[09:40] <Kamion> I think it
[09:40] <Kamion> 's reduced, though
[09:41] <mdz> elmo: gnutls11 is built, but its binaries are in universe (though depended upon by packages in main)
[09:41] <elmo> mdz: yeah
[09:43] <seb128> mdz: libxml++2.6 is new in deb and not in universe ... what should be done to get it, just ask a sync ?
[09:43] <elmo> libgcrypt11 would be another one to merge
[09:44] <elmo> seb128: we're missing new packages ATM, I'll hopefully fix that tonight
[09:44] <seb128> ok, thanks
[09:45] <mdz> elmo: libgcrypt11 is safe to overwrite with the version from Debian, please sync it
[09:45] <elmo> cool, done
[09:46] <elmo> then I can move it to main ;-)
[09:56] <Keybuk> \_ patch -stuN -z .magic-orig -r /home/scott/magic.patch-reject -p1
[09:56] <Keybuk> I love that command
[09:56] <Keybuk> "set patchers to stun"
[09:58] <mdz> Keybuk: you so added the -u just to make a word :-P
[10:11] <graham> tseng, are you the tseng of mono repository fame?
[10:11] <tseng> sure
[10:11] <graham> would you be interested another package? I've recently packaged Spam Trainer.
[10:12] <tseng> its not very useful to me
[10:12] <tseng> but if it works, you can email it to me and ill upload it eventually
[10:12] <graham> fair enough
[10:12] <elmo> WTF is with bugzilla and firefox?
[10:13] <graham> incidentally, I had an error trying to build tomboy myself from the source package and I think it might be missing a build dep.
[10:13] <elmo> Kamion: ?
[10:13] <tseng> graham: there is a syntax error
[10:13] <tseng> missing )
[10:13] <graham> I think it was the perl XML libs, but I've not got my laptop running right now and might be wrong. I'll check for the ) though.
[10:13] <graham> do you use pbuilder?
[10:13] <tseng> i use dpkg-buildpkg
[10:14] <graham> okay. I'll check it out when I get my laptop out as I need to run spam trainer through pbuilder anyway.
[10:15] <tseng> you can fix the package by fixing the ) in debian/control
[10:15] <graham> ok, thanks.
[10:15] <tseng> nps
[10:17] <amu> re
[10:18] <amu> mark#s talk was nice, and ubuntu get into the final for an award  
[10:19] <mdz_> elmo: which bugzilla/firefox bit in particular?
[10:19] <amu> sabdfl: wb 
[10:19] <mdz_> elmo: you can file those bugs in bugzilla, some of them are already there and some have been fixed
[10:19] <elmo> mdz: it keeps hanging when I scroll down with the wheely mouse button
[10:19] <sabdfl> hey all
[10:20] <mdz_> elmo: ah, that one
[10:20] <sabdfl> hey amu, long time no see :-)
[10:20] <mdz_> elmo: happens with some other sites, too
[10:20] <sabdfl> lamont: long live the...?
[10:20] <sabdfl> amd64
[10:20] <lamont> and we have buildd/sbuild for amd64 architectures now.
[10:20] <mdz_> elmo: https://bugzilla.ubuntu.com/show_bug.cgi?id=1962
[10:20] <amu> sabdfl: i joined the social event, ubuntu was was one of the final getting an award   
[10:21] <sabdfl> cool!
[10:21] <lamont> hoary.all.amd64:Total 858 package(s) in state Needs-Build.
[10:21] <lamont> hoary.all.i386:Total 1 package(s) in state Needs-Build.
[10:21] <lamont> hoary.all.powerpc:Total 123 package(s) in state Needs-Build.
[10:21] <sabdfl> who was issuing th awards?
[10:21] <lamont> amu: cool!
[10:21] <sabdfl> "best nekkid people"?
[10:21] <amu> sabdfl: LWE for best distros's best database ..... 
[10:21] <mdz> who was _receiving_ the award? :-)
[10:21] <amu> skole :) 
[10:21] <Keybuk> elmo: where do you want a list of packages we can sync from Debian?
[10:21] <Keybuk> (because they took all our changes)
[10:22] <sabdfl> *cough*
[10:22] <sabdfl> que?
[10:22] <mdz> Keybuk: yes, he does :-)
[10:22] <mdz> amu: surely you're joking, Mr. Mueller
[10:22] <elmo> keybuk: mail or url link - whatever
[10:22] <sabdfl> skole won an award, over ubuntu?
[10:22] <amu> mdz: it's like this 
[10:23] <Keybuk> elmo: http://people.ubuntu.com/~scott/revert
[10:23] <Keybuk> that's the first chunk
[10:23] <amu> guess they should add new new award, for the best debian distro
[10:24] <mdz> amu: confused
[10:24] <mdz> who gave an award to which distribution, and who accepted it?
[10:24] <amu> mdz: ... the last years all time debian won 
[10:24] <elmo> Keybuk: you're 100% sure about this?  i can't undo once I start
[10:24] <elmo> (sanely anyway)
[10:24] <amu> LWE gives some awards to the projects.  
[10:24] <lamont> elmo: and insanity takes is _soo_ painful.
[10:25] <dante_> pardon ... could someone point me to the correct 'registration' page in the official ubuntu site? I see the 'log in' button but can not find an initial 'registration' area.
[10:25] <amu> the distro award goes to skole this year 
[10:25] <Keybuk> elmo: yeah, they all resulted in a zero debian->hoary diff
[10:25] <Keybuk> (well, just changelog mogs)
[10:25] <mdz> amu: bu tyou said ubuntu received an award?
[10:25] <mdz> s/bu ty/but y/
[10:26] <mdz> Keybuk: did you eyeball them?
[10:26] <Keybuk> mdz: yup
[10:26] <Kamion> elmo: yep?
[10:26] <amu> mdz: one of the final canditates ... but skole won at least 
[10:26] <mdz> amu: ah, ok. so ubuntu was a finalist
[10:26] <amu> mdz: yap 
[10:26] <mdz> but did not receive an award
[10:27] <amu> yap 
[10:27] <elmo> Keybuk: ok, doing now
[10:28] <graham> it appears as though "skjermbilder" means "screenshot" in norwegian. neat. I've never heard of skoke before; what did they win for?
[10:28] <__daniel> graham: good question :-)
[10:28] <graham> I'm thinking it must be amazing...
[10:29] <Kamion> skolelinux are a Norwegian Debian derivative; they do mass rollouts of preconfigured systems with automatic LDAP setup etc. in schools.
[10:29] <Kamion> quite specialised
[10:30] <Kamion> any udev experts in the house?
[10:30] <graham> thanks
[10:30] <Kamion> how do I make it create /dev/fb/0?
[10:31] <Kamion> I'm using devfs.rules and compat-full.rules
[10:31] <Keybuk> what's it creating currently?
[10:32] <__daniel> Kamion: this is what i got in devfs.rules:  KERNEL="fb[0-9] *",  NAME="fb/%n",
[10:32] <Kamion> __daniel: yes, likewise
[10:32] <Kamion> Keybuk: nothing for the framebuffer device that I can find
[10:32] <Keybuk> Kamion: do you have the appropriate module for it loaded?
[10:32] <Kamion> Keybuk: yes
[10:32] <Kamion> vga16fb and fbcon
[10:33] <Keybuk> and you're not getting either /dev/fb0 or /dev/fb/0 ?
[10:33] <Kamion> I can make everything work if I create the device node by hand
[10:33] <Kamion> Keybuk: nope
[10:33] <elmo> Keybuk: k, imported
[10:33] <Kamion> it's possible hotplug was hosed actually, I'll try again
[10:34] <Kamion> minor detail of having the wrong thing in /proc/sys/kernel/hotplug ...
[10:34] <Keybuk> elmo: imported what where?
[10:35] <__daniel> Kamion: works at my place... ah ok
[10:35] <Kamion> __daniel: this is in the installer FWIW, it's a very minimal system and entirely possible I've left bits out :)
[10:35] <Kamion> trying to figure out the chain of code involved
[10:36] <__daniel> Kamion: oh... i see
[10:36] <Keybuk> 98819394 bytes transferred in 2 seconds (62.16M/s)
[10:36] <Keybuk> Total 203 files transferred
[10:36] <Keybuk> whee
[10:39] <Keybuk> let's see if I get any messages
[10:49] <Keybuk> Kamion: I don't think he did import my key
[10:49] <Keybuk> no response yet
[10:50] <elmo> no, not that
[10:50] <elmo> the all-merged-up stuff is imported from sid
[10:50] <Kamion> ah
[10:50] <Kamion> sorry, misinterpreted
[10:51] <Kamion> grr. There's a delay between a module being loaded and the device node being visible, isn't there?
[10:51] <Keybuk> elmo: see, now you have 203 files in the queue to get rid of as well as import my key <g>
[10:51] <Keybuk> Kamion: yeah, but usually not a huge one
[10:52] <Kamion> Keybuk: is it predictable in any way?
[10:52] <Keybuk> no more than a second or two
[10:52] <Kamion> I don't want to insert random sleeps, that sucks
[10:52] <Keybuk> Kamion: no, it's basically just how long it takes for the stuff in /sys to appear
[10:52] <Kamion> is there any way to say "wait until everything's synced"?
[10:52] <Keybuk> hotplug is full of "wait for file to appear" checks
[10:52] <Kamion> that's really crap
[10:52] <Keybuk> for which?
[10:53] <Keybuk> if you're doing a hotplug event, things will be synced at that point
[10:53] <Keybuk> (aiui)
[10:53] <Kamion> modprobe fbcon -> wait for /dev/fb/0 to show up
[10:53] <Kamion> in a sane world you can expect it to be there from the point modprobe exits
[10:54] <Kamion> it's not good for the installer to sit around thinking for ages, especially since depending on the system (e.g. serial consoles) /dev/fb/0 might never appear
[10:54] <Keybuk> that'd mean it'd all have to happen in kernel space
[10:54] <Kamion> I don't care how it's implemented :)
[10:54] <Kamion> I just want a supposed upgrade not to be a regression
[10:54] <Keybuk> wouldn't doing the modprobe populate something under /sys ?
[10:55] <Kamion> immediately?
[10:55] <Keybuk> no :p
[10:55] <Kamion> also, sleeps are a totally obvious race condition
[10:55] <Kamion> d-i runs on some very slow hardware
[10:56] <Keybuk> the trouble is you're doing it backwards, really
[10:57] <Kamion> oh?
[10:57] <Keybuk> you're not supposed to load a module to find out if the hardware exists
[10:57] <Keybuk> you're supposed to load the module once you know the hardware exists
[10:57] <Kamion> is there a way to test for available framebuffer hardware, then?
[10:57] <Kamion> reliably?
[10:58] <Keybuk> (this obviously isn't quite finished yet, as things like ppp_generic/loop go phooey)
[10:58] <Keybuk> well, the graphics card's pci id will map to the appropriate frame buffer
[10:58] <Kamion> all I want, really, is a wait_for_everything_to_sync call
[10:59] <Keybuk> radeonfb             0x00001002 0x00005961 ...
[10:59] <Keybuk> for example
[10:59] <Kamion> sounds a bit challenging to parse at this stage, but I guess
[11:01] <Kamion> ah well, that will come later, sleep it is for now
[11:02] <Kamion> thanks for guidance :)
[11:02] <Keybuk> nite :)
[11:02] <__daniel> sleep tight, kamion
[11:02] <Kamion> sleep> i.e. /bin/sleep not bed
[11:02] <Keybuk> oh, lol
[11:03] <Keybuk> my brain saw "sleep" and thought "mmm... bed"
[11:03] <__daniel> ;-)
[11:03] <__daniel> me too
[11:04] <Keybuk> dunno why, but I've been up at 5am every morning this week
[11:42] <Kamion> ROCK
[11:42] <Kamion> udev/hotplug d-i works
[11:42] <Kamion> though I haven't tried to strip out discover yet
[11:47] <jdub> arh, no keybuk
[11:48] <jdub> what's the best way for me to upload a modified/updated gdm?
[11:48] <jdub> just grab the new release, apply changes and go?
[11:49] <seb128> good luck
[11:49] <seb128> I've tried once
[11:50] <seb128> neuro has all his changes mixed in the diff.gz without any documentation
[11:50] <jdub> yeah
[11:50] <jdub> sucks
[11:50] <seb128> Np237 took a whole week to repackage it with cdbs last year
[11:50] <jdub> i'm tempted to fork it ;)
[11:50] <seb128> use Np237's work if you do this
[11:50] <seb128> he made a good package using cdbs with all the patches splitted in debian/patches
[11:52] <seb128> jdub: hum, we want glib/gtk 2.5 in hoary now, right ?
[11:52] <jdub> sure!
[11:53] <seb128> ok, that's a bit late now to start packaging them, but that's on my todo list for tomorrow :)
[11:55] <seb128> jdub: if you want the gdm package made by Josselin you can find it in the pkg-gnome SVN (/packages/gdm)