=== __daniel [~daniel@td9091b0b.pool.terralink.de] has joined #ubuntu-devel
pittimdz: ^12:00
pittimdz: but I think since we don't have NMUs in Ubuntu, this scheme should not clash with regular uploads12:00
mdzpitti: this is not an easy question12:00
mdz-2ubuntu0 << -2woody1, for example12:01
mdzwhich is what the Debian security team would use12:01
pittimdz: right, but we don't use woody version numbers, do we?12:01
pittimdz: well, _if_ we update a woody version, then we need to think about a different scheme12:01
=== pitti shakes
pittiwould be consistent, but ugly12:02
pittibut I think that won't happen very often12:02
=== lamont_r [~lamont@209-181-74-164.dnvr.qwest.net] has joined #ubuntu-devel
pittimdz: okay, I uploaded glibc (#2743) and imagemagick (#2769)12:04
pittimdz: gee, we urgently need version (or at least distribution) tracking in our bts12:07
mdzpitti: you are learning the pain of security :-)12:07
=== lamont_r chuckles
pittimdz: I left the bug open for now (normal/target Hoary) until I upload them to hoary12:07
pittimdz: but I think I deal with Hoary when I fixed all other security bugs12:08
pittimdz: that okay for you?12:08
=== rjb [~arjaybe@] has left #ubuntu-devel ["Kopete]
mdzpitti: yes12:10
mdzpitti: I highly recommend keeping your own todo list12:10
mdzseparate from bugzilla12:10
pittimdz: in any case12:10
mdzpitti: fortunately sarge sorts below ubuntu :-)12:11
pittimdz: and etch, too :-)12:11
mdzit should become part of our automated upgrade testing to ensure that this doesn't occur12:12
pittimdz: and at the time etch is actually released, I'm grandfather12:12
=== pitti hides
pittimdz: you mean it does not occur that our updates sort below the current Debian stable?12:12
mdzpitti: not only the current Debian stable, but anything that we support upgrades from12:12
mdzwoody->warty is a supported upgrade path12:13
mdzso warty must be > woody12:13
lamont_rmdz: I thought sarge would do 1.sarge...12:13
mdzI believe we checked at one point12:13
mdzand all of the packages in warty/main had higher version numbers than woody12:13
mdzwhich is fortunate12:13
lamont_rmdz: that won't always be true, you know...12:14
mdzlamont_r: right, but we should ensure that it is true at release time12:14
mdzuntil we fix this awful version number problem once and for all12:14
pittimdz: oh, gaim was missing12:15
pittimdz: according to my log, I already uploaded gaim to the security queue, but it is not yet in the security packages file12:16
lamont_rmdz: Just thinking of the case where sarge happens to release just before the grumpy upstream version freeze, or some such.. :-(12:16
pittimdz: I just checked all my logs, gaim is the only security upload I forgot about in above list12:17
mdzpitti: the security queue is different from the normal upload queue12:17
mdzpitti: packages do not automatically go into the archive via security12:17
mdzthey must be processed by hand12:17
pittimdz: yes, that's why I'm asking whether it is still present12:17
mdzgaim                        | 1:1.0.0-1ubuntu1.1    | source amd64 i386 powerpc     | 2 days old12:17
pittimdz: or whether I somehow forgot to actually upload12:17
pittiah, okay12:18
=== Ferry [~Ferry@cust.15.118.adsl.cistron.nl] has joined #ubuntu-devel
pittimdz: I prepared the first USN for gs-common, of course the package list is still missing.12:22
pittimdz: http://www.piware.de/usn-gs-common.txt , could you please have a short look? 12:22
mdzpitti: great12:22
mdzpitti: the date should be right-justified12:22
mdzbut I think amber should do that for you now12:23
pittimdz: oh, amber also creates this template?12:23
pittimdz: I just copied the template from the website, where it is not justified at all12:23
pittimdz: fixed12:23
mdzpitti: yes, amber will give you a template to be filled in12:24
pittivery nice12:24
mdzthe website stuff is cut-and-waste from the emails I sent12:24
pittiI prepare the stuff for the other packages and send them out as soon as I get the amber mail12:24
pittiProbably not today any more, I'm exhausted and it's late12:25
pittiand libc6 will take a while to build (took 4 hours on my machine)12:25
elmopitti: 30 mins in the DC :)12:25
pittielmo: okay, okay, you have better machines to play with :-)12:26
jbaileypitti: How many passes is that?12:26
elmohmm, actually, it's 30 i386, 43 powerpc, 48 amd6412:26
pittijbailey: pardon?12:26
elmopitti: how many different versions of glibc do you build for giggles12:27
jbaileypitti: processor/thread combinations.  In Debian on ia32, we usually have 3 (386-linuxthreads, 386-nptl, 686-nptl)12:27
pittiI just tried it on my desktop (686)12:27
pittibut I thought debuild would build them all?12:28
jbailey4 hours just seems awful high.12:28
pittiI remember seeing much of the build log parts more than once12:28
elmojbailey: it's stock Debian source, so it'll be whatever debian does for 38612:28
elmos/stock/pretty-much \1/12:28
pittijbailey: I don't have the fastest machine, though :_)12:29
=== azeem once compiled glibc on 400 MHz notebook for Debian GNU/Hurd. It sucked.
mdzpitti: since we only upload source, it is important to test the binaries which are built by the buildds12:31
mdzat least one arch12:31
jbaileyazeem: Debian GNU/Hurd takes about 4 times as long in my experience.  I usually leave 3 hours on my 1.8ghz p4.12:31
pittimdz: 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)12:31
mdzpitti: hm?12:32
jdubyo jbailey 12:32
mdzpitti: the binaries built by the buildds12:32
azeemjbailey: it definetely wasn't 12 hours on my 1.8ghz p412:32
azeem(on Hurd)12:32
jbaileyg'morning Jeff. =)12:32
pittimdz: ah, I see12:32
azeemmore like 5 or 612:32
pittimdz: sorry, confused12:32
jbaileyazeem: We do 3 passes on i386, though.  On the Hurd we only do one.12:32
azeemah :)12:32
jdubmdz: how'd the meeting go? i fell asleep late last night putting pipka to bed12:46
=== doko [doko@dsl-084-057-046-053.arcor-ip.net] has joined #ubuntu-devel
seb128hello jdub 12:46
jdubpants off seb128 12:46
mdzjdub: which one? CC?12:46
mdzit went pretty well, I was in and out a bit at the end12:46
mdzjdub: mako has transcript and is preparing a summary12:46
makojdub, mdz: am planning on finishing that one tonight12:50
pittimdz: I have all USNs ready. I need some sleep, I will send them out tomorrow (when the amber mails hopefully arrived)01:05
pittiGood night everybody!01:05
mdzpitti: wait01:07
pittimdz: still here01:07
mdzpitti: the amber mails will not arrive until I run amber01:07
mdzand the advisories should go out immediately after that01:07
mdzglibc does not seem to be built yet01:07
jdubdaniels: around?01:08
pittihmm, is there an approximate ETA?01:08
pittimdz: I can stay up for another hour, if necessary01:08
mdzpitti: elmo is making some adjustments to amber01:12
mdzpitti: which ones are you ready to send out?01:12
pittimdz: all but gaim01:12
pittimdz: I'm currently typing gaim01:12
mdzpitti: ok. I have tested gaim and it works fine; I can't test MSN but I see no regressions in AIM/jabber01:12
mdzpitti: which ones are the simple temporary file stuff?01:13
mdzgs-common, glibc, gettext....?01:13
pittimdz: gettext, glibc, gs-common, postgresql01:13
pittimdz: do you have the two CC'ed mailing list adresses at hand?01:13
mdzpitti: gaim, gettext, gs-common, postgresql are built and ready01:14
pittimdz: please give me some more minutes01:14
mdzglibc and imagemagick are not built yet01:14
pittimdz: I need to change my subscription to bugtraq01:14
pittimdz: I subscribed as martin@piware.de, but I need my canonical address01:15
Kamionmdz: I'm producing the required groff patch now, FYI01:15
mdzKamion: thanks, keep pitti informed01:15
mdzpitti: bugtraq@securityfocus.com, full-disclosure@lists.netsys.com01:15
pittimdz: do I need to subscribe full-disclosure to post to it?01:16
mdzpitti: I am not sure01:16
mdzpitti: you need to be subscribed anyway01:16
pittimdz: okay01:17
danielsjdub: sup?01:17
=== nmf [~nmf@] has joined #ubuntu-devel
jdubdaniels: do you think composite-on-by-default is likely for hoary?01:18
danielsjdub: we can try it, but I'm uncomfortable doing it without a fair bit of testing01:19
danielsbut if we test it in the hoary tree and it's not too bad, we can keep it on for main01:19
jdubthough it's going to be different for different video cards, right?01:20
danielsdue to render accel01:22
danielsi think radeon's is pretty solid, and that in the nvidia binary driver can be a bit odd sometimes01:22
lifelessis it easy to turn on/off ? I.e. a control panel ?01:22
danielslifeless: Section "Extensions"\n\tOption\t"Composite"\t"Enabled"\nEndSection01:22
danielsit's not configurable on-the-fly01:23
lifelessah. perhaps that should be a goal - an option for folk to switch, but htey have to log out and in01:23
mdzpitti: are you ready to publish?01:23
mdzpitti: elmo is finished and I am ready to amber01:23
pittimdz: I don't get a confirmation email from netsys01:23
pittimdz: I subscribed several minutes ago...01:24
danielslifeless: mmm01:24
pittimdz: otherwise, bugtraq subscription is okay, all USNs are ready01:24
mdzpitti: the important thing is that it goes to ubuntu-security-announce and the webiste01:24
danielslifeless: 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
pittimdz: I can send the full-disclosure mail later if it has member-only posting 01:24
danielss/don't want/want/01:24
mdzpitti: have you read the advisory guidelines I put in the wiki?01:25
pittimdz: not yet01:25
mdzpitti: http://wiki.ubuntu.com/SecurityNotification01:25
pittimdz: ah, I read this01:26
mdzpitti: for the temporary file vulnerabilities, the advisory should indicate with what privileges the overwriting takes place01:26
mdzif it is the privileges of a user invoking a program, or a system process, etc.01:26
mdzbecause this makes a big difference in the impact01:26
elmohmm, lamont has the buildds set to take too much01:26
pittimdz: hmm, I have to add this01:27
elmowell, 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 less01:27
mdzpitti: let me know when you are ready for me to review the text01:27
mdzelmo: what are we going to do about getting new packages from sid into hoary/universe?01:28
mdz(rather than new versions)01:28
mdzor are you already doing that?01:28
elmoI'm going to deal with that in a bit.. also have to deal with removed packages, which is alittle more fun01:28
elmosince we have no record of where non-Debian stuff in universe came from (whee)01:28
elmocan someone prioitize merging/clearing-for-sync libldap2 and any other gnutls using libraries?01:29
elmowithout that hoary's going to be severely broken dep-wise01:29
elmomdz: btw, I assume I'm okay that people don't need to bother you/jeff for clear-to-sync mails?01:30
Kamionmdz: ok, uploaded to Debian, cced pitti on my mail to upstream01:30
mdzelmo: correct01:31
mdzelmo: dude, the merges are done01:32
mdzKeybuk said they were finished this afternoon01:33
=== elmo hails keybuk's turing complete merge-meister
KamionI guess that would explain why I still have the majority of uploads to hoary :P01:34
pittimdz: ready: http://www.piware.de/usn/01:39
pittimdz: I hope my English is not too bad at this time01:39
=== lamont_r [~lamont@209-181-74-164.dnvr.qwest.net] has joined #ubuntu-devel
pittimdz: could you do some proofreading, please? For helping a USN author newbie? :-)01:39
elmoRecently, Trustix Secure Linux discussed a vulnerabily in the01:41
=== daniels stares at elmo.
elmo way, which allowed a symlink attack for overwriting01:42
pittielmo: thanks, fixed01:42
elmoI'd say: way which allowed a symlink attack to overwrite01:42
elmopersonally, but that's possibly just me01:42
=== __daniel [~daniel@td9091b0b.pool.terralink.de] has left #ubuntu-devel ["Verlassend"]
pittielmo: I take the word of a native speaker01:42
elmosame for gs-common then01:43
pittielmo: fixed both errors in all USNs01:44
pittiand uploaded again01:44
elmoA buffer overflow and two remote crashes were recently discovered in01:44
elmothe MSN protocol handler.01:44
elmoI'd say "gaim's MSN protocol handler"01:44
elmootherwise it sounds like MSN is buggy, IYSWIM01:44
mdzelmo: which is not altogether inaccurate, but irrelevant :-)01:45
mdzpitti: reading now01:45
elmopitti: still misspelling vulnerability in postgres01:45
mdzpitti: have you verified the CVE names against the CVE website?01:45
pittimdz: yes01:45
pittielmo: spell fixed01:46
elmo"Since imagemagick might be used in custom printing systems," <- s/might/can/ is more idiomatic and solves the duplication in the next sentence01:47
mdzpitti: it is a good idea to run an aspell check on all of the text01:50
pittiwill do next time01:50
mdzpitti: s/files of/files with the privileges of/01:51
mdzpitti: in at least gettext01:51
mdzthe files would not need to belong to that user, only be writable by them01:52
mdzpitti: also s/overwrite/create or overwrite/01:52
=== __randy__ [~randy@sclab-25-433.sclab.clarkson.edu] has joined #ubuntu-devel
pittimdz: fixed01:55
tsengcool, jimmac is using ubuntu01:56
mdzpitti: looks good01:56
mdzpitti: ready to release?01:57
=== __randy__ [~randy@sclab-25-433.sclab.clarkson.edu] has joined #ubuntu-devel
pittithumbs up! I even got my f-d subscription01:57
elmohey, that's a point01:57
elmowe need a "who uses ubuntu" section on the website01:57
jduboh dudes01:57
jdub^ jimmac likes Ubuntu on his Mac01:57
tsengoh dudes, i just said that !01:57
mdzwho is jimmac?01:57
danielselmo: that'd be easy01:57
mdzI assume he works for ximian01:57
danielselmo: 'all the cool kids'01:57
jdubmdz: jakub steiner, gnome icons man01:57
danielsmdz: yeah, he's one of their crazy art guys01:58
mdzelmo: amber blew up01:58
jduband ximian/novell art biatch01:58
mdzelmo: jabbered you the output01:58
elmono you haven't?01:58
mdzlike hell I haven't01:58
elmohave NOT01:59
mdzsent again?01:59
mdzperhaps you are not logged into jabber :-P01:59
elmoI so am dude01:59
mdzooh, jabber is broken01:59
elmogeez.. first he insists he's sent me mail.. then he insists he's sent me jabbers..01:59
mdzI disconnected from jabber and tried to reconnect, and it hangs02:00
jdubdavyd's latest sshot has 'sudo suspend' ;)02:02
jdubmdz: btw, why did we go for "add a user line to sudoers" instead of a wheel-group equivalent set up?02:02
mdzjdub: no one even mentioned the possibility, including you :-P02:03
mdzcould very well be simpler to do it that way02:03
tseng%wheel  ALL=(ALL)   ALL02:03
jdubmdz: btw, http://cipherfunk.org/diary/archives/monthly/2004-10.html#e2004-10-26T18_44_02.htm02:03
jdubmdz: i thought of it a few times, but stupidly didn't mention it, because i thought it had been passed over.02:04
jdubmdz: fwiw, that's how OS X does it02:04
jdubmdz: would it be a PITA to transition to it?02:04
mdzjdub: not for new installs02:04
tsenghttp://ronald.bitfreak.net/images/g-v-a.png <- whats different in this shot?02:04
jdubi guess it wouldn't for upgrades either; you'd only have to add one line02:05
jduband existing users already have sudo love02:05
jdubtseng: dunno02:05
tsenglooks the same old to me =/02:05
jdubtseng: although it has been rewritten02:05
=== mjg59 fails to get ACPI on the craptop working under FreeBSD
jdubmdz: interesting patches02:06
jdub(paul is 'ultrafunk' aroudn irc)02:07
jdubyou know02:08
jdubi was thinking last night02:08
mdzjdub: what did he have on the laptop before?02:08
jdubprobably fedora02:09
Kamionmdz: I saw it mentioned a number of times, actually02:09
jdubwe make a lot of decisions on irc02:09
jduband talk about interesting stuff here02:09
Kamionmdz: always when more important stuff was happening, though ...02:09
jdubwhile some people may only be able to track ubuntu-devel02:09
mdzKamion: I remember it was mentioned late in the release cycle02:09
jdubi want to make more of an effort to use that list02:09
mdzbut not back in Oxford when we made the change02:09
Kamionmdz: *nod*02:09
Kamionmdz: I think it'd be good for new installs, but I was a bit scared to use "wheel"; that group has a lot of baggage02:10
mdzKamion: agree02:10
Kamionpeople rather expect it to be gid 0, too02:10
elmogod damn it, I suck.02:10
mdzpitti: still with us?02:10
pittimdz: yes02:10
Kamionadding the initial user to gid 0 would suck :-)02:10
mdzpitti: minor train wreck in generating the advisory template02:11
=== lamont_r [~lamont@209-181-74-164.dnvr.qwest.net] has joined #ubuntu-devel
Kamionanyway, this alleged going-to-bed thing ...02:11
jdubadduser jdub sudo ;)02:11
danielsKamion: ... is for the weak!02:11
pittimdz: in the meantime I created all mails and put them into my outgoing queue; so I just need to add the package lists02:11
=== daniels heads out of range of WiFi.
KamionI'll be out for a fair bit of tomorrow; picking up the girlfriend from visiting family an hour or so's drive away02:12
mdzKamion: ok, enjoy02:12
danielsenjoy :)02:12
Kamionwill catch up as per usual :)02:12
elmokatie@jackass:/srv/ftp.no-name-yet.com/queue/accepted $ amber 3-1 gs-common_0.3.6ubuntu1.1*.changes -n | less02:15
elmoamber is only useful for members of the security team, you tramp.02:15
Kamionyou can smell elmo-written scripts a mile away02:15
=== __randy__ [~randy@sclab-25-433.sclab.clarkson.edu] has left #ubuntu-devel ["Leaving"]
jdubthey smell like schhhnnaaaakke02:18
=== shlomil [~shlomi@] has joined #ubuntu-devel
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
=== lupus_ [~lupus@kn-ivl-2.kuleuven.net] has joined #ubuntu-devel
pittiHi sjoerd! nightshift, too?02:34
mdzpitti: you should have an email with the gs-common template now02:34
pittimdz: not yet02:34
elmoOct 27 01:34:10 fiordland postfix/smtp[10351] : D3041B68002: to=<mdz@alcor.net>, orig_to=<security@ubuntulinux.org>, relay=redir-mail-telehouse1.gandi.net02:35
elmo[] , delay=2, status=sent (250 Ok: queued as ED36E445C1D)02:35
elmoOct 27 01:34:39 fiordland postfix/smtp[10352] : D3041B68002: to=<martin@piware.de>, orig_to=<security@ubuntulinux.org>, relay=mail.piware.de[] 02:35
elmo, delay=31, status=sent (250 OK id=1CMblu-0004Rv-VO)02:35
mdzit liked me better02:35
pittimdz: now it's there02:35
pittiprobably took a while to get though SA, procmail, and the like02:36
mdzpitti: gettext inbound02:36
mdzand postgresql02:36
mdzand gaim02:37
elmohmm, glibc's taking too long, lemme check on it02:37
pittimdz: the template shows all generated debs; shall I cut that down to the really affected ones?02:38
elmook, we need lamont for glibc02:38
mdzhmm, he was just here02:38
elmothe buildd mail config is broken for the powerpc and amd64 buildds it built on02:38
mdzcan we retry it and hope it gets a different one?02:38
mdzsorry, channeling lamont there02:39
mdzI think I am going to need to go buy some sandbags and build a dike02:39
mdzbefore I drown02:39
pittimdz: ready to fire the first USN (gs-common)02:40
mdzpitti: whenever you are ready, or if you want me to look at it first, let me kno02:41
pittimdz: it's just the known notice with the package list appended02:41
pittito: u-s-a, cc: full disclosure, bugtraq, signed02:41
mdzpitti: go ahead02:41
pittimdz: for gettext, only the really affected "gettext" or shall I include all other packages (gettext-base etc.) as well?02:44
mdzpitti: if you know that the bug is only in one binary package, you can list only that one02:44
pittimdz: as I already did, okay02:45
pittimdz: so I can also remove the other packages/md5sums/etc?02:45
mdzpitti: leavae the md5sums02:45
pittigettext is away02:46
pittiI think I should wait for the arrival of gs-common before sending further mails...02:46
mdzI approved the list messages02:46
pittiargh, why does my post to u-s-a await moderator approval???02:46
mdzpitti: every post is moderated02:46
pittiah, okay. So no error on my side02:47
=== Nigelenki [~bluefox@pcp485126pcs.whtmrs01.md.comcast.net] has joined #ubuntu-devel
mdzpitti: trying to reach lamont02:50
=== Nigelenki [~bluefox@pcp485126pcs.whtmrs01.md.comcast.net] has joined #ubuntu-devel
mdzhe is enroute02:51
pittiah, the first two mails came through :-)02:51
=== pitti is excited
mdzpitti: :-)02:53
pittimdz: I sent the three other USNs (all except libc)02:53
mdzpitti: do you have privileges to add them to the website?02:54
pittimdz: time to try out my password :-)02:54
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
elmoyou guys don't need me for anything else do you?02:57
=== Nigelenki [~bluefox@pcp485126pcs.whtmrs01.md.comcast.net] has joined #ubuntu-devel
pittielmo: thanks a lot for your help!02:57
pittimdz: it seems as if I could edit the page02:57
mdzelmo: all the ambering has worked fine, so I don't anticipate a problem with glibc, thanks02:57
=== lamont_r [~lamont@209-181-74-164.dnvr.qwest.net] has joined #ubuntu-devel
elmowell, mobile's on anyway, if something comes up.. 02:58
elmonight all02:58
jdubnight elmo02:58
lamont_rstill here?02:58
pittimdz: when I click on "add error reference" on http://www.ubuntulinux.org/support/documentation/usn/, I get a permission denied02:59
pittimdz: is that the right way to add another USN?02:59
elmolamont: wha?02:59
lamont_rmdz said I needed to poke a couple buildd's... wondering which machines...03:00
elmowhatever built glibc03:00
elmoadare and ..03:00
lamont_rah, yes03:00
elmothe glibc build log bounced, AFAICT03:01
=== lamont_r goes to deal with it
elmodo I want to know why, or will it make me cry?03:01
elmothe size limit seems to be set to 0..03:01
lamont_rmessage_size_limit, iirc03:01
mdzpitti: yes, that's right03:01
mdzpitti: you must not have permission03:01
lamont_rdunno - I'll get 'em signed/uploaded, and triage/fix/email you when I get home03:02
pittimdz: can I get it?03:02
elmolamont: k03:02
mdzpitti: I don't know how to do that, but I'll look03:02
mdzpitti: I don't see a way to do it; I may not have permission03:04
pittimdz: hmm; Alexander Limi and Jane should know how?03:05
pittimdz: maybe we can sort that out tomorrow and you copy the USNs to the page?03:05
mdzpitti: yes, I will need to do that03:05
pittimdz: sorry :-/ I will ask somebody tomorrow to give me permissions03:06
lamont_rpitti: glibc failed on amd6403:06
lamont_ramd64 not in arch list. :-(03:06
lamont_rmissed it by that much...03:06
pittilamont_r: I only changed a script, after all...03:06
elmoso, err, yeah, how did it build last time?03:07
pittilamont_r: does the Debian version fail as well?03:07
lamont_rcould, dunno03:07
lamont_relmo: 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_ristr adding a touch to debian/rules for warty...03:08
lamont_rpitti: bbiab03:08
mdzwait, what?03:10
mdzdoes someone have the failed logs in hand?03:13
mdzand is that same someone working to fix the problem?03:13
elmoI'm not working on it tho03:17
=== bluefoxicy [~bluefox@pcp485126pcs.whtmrs01.md.comcast.net] has joined #ubuntu-devel
mdzpitti: all USNs should be published on the website now03:17
pittimdz: nice, thanks03:18
pittimdz: you did not include the pacakge md5sum lists, as you did for your first two USNs?03:19
mdzpitti: yes, I will remove them from the first two as well03:19
jdubmdz: btw, you don't have to turn emergency moderation on, or list yourself as admin and moderator03:19
pittimdz: probably nicer03:19
mdzjdub: the other half of that sentence would have told me the alternative method to get what I want :-)03:20
jdubmdz: it was already set up correctly03:20
mdzjdub: it was set up for every message to be moderated?03:20
jdubnew members are moderated by default03:20
jdubon that list03:20
mdznew members?03:20
jdub"By default, should new list member postings be moderated?" -> YES03:21
mdzwhere is that set?03:21
jdubsender filters page under privacy03:21
mdzfirst place I would have looked03:22
mdzwhat's a "new list member"?03:22
mdzah, I see03:23
mdzso when new users subscribe, they get their individual moderator flag checked03:23
mdzI think I feel better with it being globally locked down03:23
pittimdz: interesting post on u-users: sb asked how to verify GPG keys on the u-s-a list03:26
pittimdz: shall we put the fingerprints on the website?03:26
mdzpitti: the same way to verify any other GPG key03:26
mdzthrough the web of trust03:26
pittimdz: still it might be a good idea to have the fingerprint on a website03:26
mdzit only encourages bad habits, really03:26
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
mdzbut we can do that03:27
pittilamont_r: already back?03:28
=== pitti nearly falls to sleep
pittimdz: 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
=== lamont_r [~lamont@209-181-74-164.dnvr.qwest.net] has joined #ubuntu-devel
mdzlamont_r: ?03:31
pittilamont_r: ah, welcome back :-)03:31
=== lamont_r screams at the laptop
lamont_rso, 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
mdzlamont_r: purge acpid?03:32
lamont_rdid I miss anything worth reading the log file for ?03:32
pittilamont_r: not really03:33
lamont_rok - I'll see it when I get home03:33
mdzlamont_r: you missed me losing my mind over how we managed to release warty with a glibc that FTBFS03:33
lamont_rwhich'll be relatively soon03:33
lamont_rmdz: I swear I didn't cheat on that one.03:33
pittilamont_r: any ETA? it's 3:34am and I can hardly keep my eyes open03:33
mdzlamont_r: so you think it's a timing thing?03:33
mdzpitti: go ahead and sleep, I will take care of it03:34
mdzor it can wait until tomorrow03:34
mdzpitti: thanks for staying up03:34
pittimdz: okay, I send you the announcement03:34
pittimdz: thanks a lot for mentoring me!03:34
lamont_rmdz: glibc has traditionally been victim to that.03:34
mdzlamont_r: we need for warty-security build failures to go to security@03:34
lamont_rnote: never patch both the source and target in the diff.gz, unless you deal with it somehow in rules...03:34
pittimdz: USN sent to you03:35
pittithanks everybody and good night!03:35
mdzpitti: night03:35
mdzlamont_r: can you fix it up with an upload to warty-security?03:35
lamont_rmdz: meaning fix and re-upload source, I assume?03:36
mdzlamont_r: correct03:36
lamont_rnp.  car about done here, then will be home relatively soon and deal with it.03:37
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
=== jamesh [~james@203-59-217-232.dyn.iinet.net.au] has joined #ubuntu-devel
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
hornbeckI am liking the new wiki :-)04:05
jdubthere is an optimisation that we MUST HAVE in hoary!!!04:11
mdzoh really???04:11
jdubit is crucial for enterprise adoption!!04:12
tsengyeah, meh04:13
mdzjdub: we should add a shutdown script which executes sync(1) in a shell loop 100,000 times04:14
mdzthen replace our sync(1) with this one and claim a performance benefit04:14
hornbeckjdub: that is nice04:14
mdzd00d3rz 1 r3c0mpi1e/> w1th -p1p3 4n/> 1t /2u/\/z 50kx f4st3r!!!1104:15
=== SuperLag [~colbyirc@CPE-69-76-188-71.kc.rr.com] has joined #ubuntu-devel
jdubmdz: that MSI bug seems to suggest that it was on and off during warty devel04:25
mdzjdub: yeah, it was supposed to end in the 'off' state04:26
=== jdub is tempted to rebuild to try, but... ugh...
jdubi don't want to build kernels anymore :-)04:26
mdzjdub: if you have a -15 or lower .deb around, you can try that04:26
mdzmaybe a preview CD04:26
mdzthe regression happened oct 1204:26
mdzsnapshot.ubuntu.com? :-)04:27
mdzfuck, raining again04:28
=== lamont is hime
lamonthome, even04:33
lamontRejected: 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.04:36
=== lamont giggles
mdzlamont: what's the word on glibc?04:42
lamontmdz: the word is that I'm home and ignoring irc while I fix it. :-)04:44
=== sjoerd [~sjoerd@fire.ipv6.luon.net] has joined #ubuntu-devel
=== robtaylor [~robtaylor@] has joined #ubuntu-devel
hornbeckhow would you make a code block in reStructuredText?04:51
mdzhornbeck: you are actually using it?04:53
mdzI found it unusable04:53
hornbeckwhat reStructuredText04:56
hornbeckit is great04:56
hornbecklike butter on your breakfast toast04:56
mdzthe first thing I tried to write, it blew up because the header markup wasn't exactly the same number of characters as the heading text05:04
hornbeckyeah, I have some problems with little lame things05:04
hornbecklike not having returns in the right places05:05
hornbeckman building beagle is a pain in the ass sometimes05:05
jdubmdz: sabdfl reckons we'll standardise on moin-style markup05:06
mdzjdub: me too05:06
hornbeckso I am learning this for nothing :-(05:06
hornbeckoh well05:06
mdzI can't believe you don't hate it :-)05:06
hornbeckI don't hate many things05:06
hornbeckI can find fun in pooping outside05:06
lamontmdz: have me remind you to beat up the person who added amd64 support 99% of the way, eh?05:07
=== lamont does a test build first
lamontsource upload should happen in about 45 minutes, give or take05:07
hornbeckis the universe repos down right now05:09
hornbeckmy apt-get just died05:09
lamonthornbeck: new day begins. :-)05:12
lamont"days" are 30 minutes long...05:12
lamontso the odds of catching a Packages.gz update or mirror pulse are greater.05:12
hornbeckI am trying to write a evolution-sharp doc, and I need all the deps to make sure I get them all05:12
lamontday begins at :03 and :33, pulse shortly thereafter05:12
hornbeckguess it will wait till tomorrow05:13
hornbeckoh so it will be back in a few?05:13
hornbeckok, I can get somethings from universe but alot of xml files are gone?05:17
hornbeckdid some stuff get taken out?05:17
hornbeckErr http://archive.ubuntu.com warty/universe libxml-namespacesupport-perl 1.08-3  404 Not Found05:18
hornbeckany ideas?05:18
mdzhornbeck: hmmm!05:20
hornbeckthat package and libxml-sax-perl05:20
hornbeckboth used to be there05:21
mdzit was moved, but not intentionally05:21
mdzI think I see what happened05:21
mdzelmo: ^^^^05:21
mdz(when he wakes up)05:21
hornbeckthey will be back correct?05:22
hornbeckcause evolution-sharp depends on them05:22
mdzhornbeck: as I said, it looks like it was unintentional05:23
hornbeckok cool05:23
mdzI'm writing an email to elmo05:23
hornbeckthank you05:23
mdzif you notice anyone else encountering that, let them know that we're aware of it and working on the problem05:25
hornbeckmdz: ok, I am off to bed anyway but was working on one more doc before bed05:26
lamontmdz: if it helps, I have a never-break-my-archive (albeit crufty) script for mirroring05:30
lamontsnags Packages.gz/Sources.gz, parses them. fetches files, installs Packages.gz Sources.gz, fills the morgue (since bandwidth hurts me.)05:31
lamontmdz: so does squirrelmail make baby jesus cry?05:32
jdubplus it uses baby-jesus-baiting php4-imap05:34
lamontany sane alternatives?05:35
jdubi haven't really found one that made me comfortable05:36
jdubatm i'm using sqwebmail at home, which has its own faults05:37
lamontthat was what led me to installing squirrelmail on a non-i386 box05:37
lamontglibc (if/when it decided to rebuild debian/control) lacked amd64 mention in control.in/libc6 :-(05:38
lamontnot sure where I got the amd64 patch when I applied05:38
=== lamont notes that killing the infinite-loop helps speed things up. 5MB of log to go
=== lamont wishes that Packages files were broken up into more bite-sized pieces
fabbionemorning guys06:29
=== SuperLag [~colbyirc@CPE-69-76-188-71.kc.rr.com] has joined #ubuntu-devel
=== SuperLag [~colbyirc@CPE-69-76-188-71.kc.rr.com] has left #ubuntu-devel ["Leaving"]
=== lamont identifies the email issue, sends elmo the info
lamontmorning fabbione06:38
fabbionelamont: how busy are adaire and yellow?06:42
fabbionei guess all the buildd are pumping 100% after the sync06:42
lamontfabbione: building hoary, basically06:42
fabbionei will soon need to start building X.org06:42
=== lamont currently has crested offline, so yellow is holding down the fort
fabbioneit compiles on i386 already06:42
lamontah, ok.  happy, happy, joy, joy06:42
fabbionelamont: i won't be that happy06:42
fabbioneit is still monolithic06:42
lamontyeah - but the breakup of the source can largely happen in parallel with people abusing the monolithic beast.06:42
lamontyou're going to break it up so that it doesn't have circular build-deps, right?06:42
=== lamont grumbles at hoary
lamontmy mirror is now out-of-date: 1353 files missing. :(06:43
=== Keybuk [scott@descent.netsplit.com] has joined #ubuntu-devel
lamontwell, 1249 when I snap debian sources over.06:44
fabbionelamont: i already did basically :-)06:44
jdubmorning Keybuk 06:49
fabbioneand i also have much less compiler warnings06:49
lamontfabbione: yeah06:49
jdublamont: i386 packages.gz for hoary looks fat! :)06:49
fabbionethan in the monolithic06:49
fabbionelamont: but it won't be uploaded until hoary release date + 106:49
Keybukjdub: morning06:49
fabbioneso everybody will live with monolithic for another release cycle06:49
jdublots of ubuntu packages in hoary ;)06:49
lamontjdub: scary thing is that universe in hoary got bigger than warty06:49
lamontditto main06:49
lamontand multiverse06:49
jdubdoesn't that make sense?06:49
Keybukjdub: I have 77 to upload at the moment, another 277-odd remaining :)06:49
jdubKeybuk: rad!06:49
jdublamont: 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
jdubmaybe we should have a jigdo equivalent for livecds06:49
jdubthat'd be rad06:49
jdubno it wouldn't06:49
jdubit'd be pretty much the same as doing a livecd build anyway06:49
=== jdub should sleep more
=== jdub is totally tempted to upgrade to hoary on his test box
fabbionelamont: i will quite soon (a couple of weeks from now) need the tools to build the livecd06:49
lamontjdub: that'll be scary06:49
lamontfabbione: np06:49
fabbionelamont: to sync the X autoconfig stuff on both live and install06:49
lamontI should have the doc done by then, esp if someone picks on me06:50
fabbioneotherwise we can find something else.. hmmmm06:51
fabbionethere is something more cool we can do...06:51
=== fabbione needs to think a bit about it
fabbionelamont: is there any simple way i can detect we are running on a LiveCD?06:57
fabbionelike a specific installed file?06:57
fabbionethat it will be there for sure 100%06:57
fabbionesince people have been asking for autoreconfig in case of a video card change06:57
fabbionethat can be done in a init script06:57
lamont /MorphixCD/  iirc06:57
fabbionethe same init script can check if we are running on a LiveCD06:57
fabbioneand instead of prompting it could reconfigure X06:57
fabbionethat will save me the time to build liveCD's06:57
lamontwe may have to actually schedule said script to run, btw.06:57
fabbioneif it needs to be done maually even better06:57
lamontyeah - telling that you're LiveCD isn't all that hard.  If it is, we'll make it not-hard. :-)06:57
fabbionei can stick it in /usr/share/xorg/reconfigure_for_live_cd06:57
fabbioneor something06:57
fabbioneand than you can force it manually06:57
lamontelmo/thom aroudn?06:57
=== lamont just screwed up, I fear
=== Keybuk feigns shock
lamontpitti around06:58
jdub184 upgraded, 4 newly installed, 6 to remove and 0 not upgraded.06:58
jdubNeed to get 69.6MB of archives.06:58
jdubAfter unpacking 7676kB disk space will be freed.06:58
jdubThe following packages will be REMOVED:06:58
jdub  alsa-base alsa-utils gdm python2.3-genetic ubuntu-base ubuntu-desktop06:58
Keybukjdub: what you upgrading to?07:00
=== lamont bets hoary
Keybuklamont: that would be a totally insane thing to do right now, surely? :o)07:00
lamontKeybuk: beyond doubt07:00
=== lamont hopes to hell elmo's scripts do what he hopes they do...
Keybuksome 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 keyring07:01
Keybuklamont: what do you hope they do?07:01
lamontput the files I just uploaded in the right place.07:01
jdubyeah, just seeing the state of the archive07:01
lamont75 seconds, and we'll know... :-(07:01
Keybukjdub: there will be a few problems where stuff sync'd from unstable depends on a package not sync'd because we modified it07:04
=== lamont bows at elmo's feet.
jdubi imagine gdm is waiting in your queue07:05
Keybukrookery scott% grep -c "^diff " results/gdm_2.6.0.4-1ubuntu1.dropped07:06
Keybukthat'll need manual review07:06
fabbioneKeybuk: if it's only question of signing and upload i can do that for you07:08
fabbioneif you think that will speed up stuff around07:09
Keybukthanks, 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 ;p07:09
lamontjdub: as for status:07:12
lamonthoary.all.amd64:Total 1599 package(s) in state Needs-Build.07:12
lamonthoary.all.i386:Total 1047 package(s) in state Needs-Build.07:12
lamonthoary.all.powerpc:Total 1354 package(s) in state Needs-Build.07:12
fabbioneKeybuk: ok07:13
KeybukXimian employee in "If somebody asks me what linux distro he should try on his mac, I'm resolutely recommending Ubuntu." shock.07:14
jdubinnit great? :)07:14
lamontnow we just need to make hoary suck less than warty, eh?07:15
=== lamont ducks
mdzKeybuk: morning07:16
hornbeckis there a bootstrap to use to setup a ubuntu chroot?07:16
lamontmdz: mcmurdo, royal, yellow.07:16
lamonthornbeck: start with debootstrap from warty07:16
Keybukmdz: morning07:17
lamontdebootstrap ... warty07:17
hornbeckgreat thanks07:17
=== Burgundavia [~Burgundav@S01060050ba84f285.gv.shawcable.net] has joined #ubuntu-devel
mdzKeybuk: what's the status of your merge project?07:18
lamontKeybuk: did we ever decide on a binNMU standard?07:18
lamontversion number, that is.07:18
Keybukmdz: 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 review07:18
mdzKeybuk: done == uploaded?07:18
Keybuklamont: no, need to get you, elmo, neuro, aj, kamion, etc. together and let you fight it out07:18
Keybukmdz: not uploaded yet07:19
mdzKeybuk: do you have a list of those which will definitely require manual review?07:19
Keybukmdz: not yet, because I hope to get the number down; it's at 251 at the moment07:20
mdzKeybuk: are you certain your key is not already in the keyring?07:21
mdzI was under the impression that it was07:21
Keybukwell, I did an upload the other day and it vanished into thin air07:22
mdzKeybuk: what package?07:22
Keybukubuntu-artwork, just before warty release07:23
mdzKeybuk: do you recall the version number?07:25
=== SuperLag [~colbyirc@CPE-69-76-188-71.kc.rr.com] has joined #ubuntu-devel
mdzthere's only one 0.2.12-1 in the log, and it was accepted07:26
Keybukyes, but Colin signed that07:27
Keybukhe had to get me out of bed because mine didn't go through07:27
BurgundaviaWho would I talk to about the expired cert on bugzilla?07:28
mdzBurgundavia: no one; it's already filed in bugzilla07:29
Burgundaviamdz: so I assume the powers that be are going to fix it?07:29
mdzBurgundavia: it's not a particularly high priority07:29
Burgundaviaok, just wondering07:30
mdzbeing able to spoof our bugzilla wouldn't be particularly rewarding :-)07:30
Burgundaviasort of ironic though07:30
fabbionepatch forwarding is basically completed07:34
fabbionemdz: ^^^07:34
fabbionemdz: + we build on i38607:34
lamontStarting proxy server: 2004/10/26 23:32:36| WARNING cache_mem is larger than total disk cache space!07:34
lamonthrm...wonder how to bump the disk cache size..07:34
=== mdz high-fives fabbione
mdzlamont: cache_dir07:35
fabbionemdz: now we need to fix all the packaging part :-)))07:37
fabbionethat it's going to be... hmm evil07:37
fabbionemdz: since we switched to monolithic i will NOT neeed root on amd64 or ppc07:39
fabbionemdz: all the build-deps should be there already07:39
mdzfabbione: what about for testing?07:39
fabbionemdz: even if i have root on our boxes, there is nobody in front of them to see what's happening on the screen07:41
fabbionei will need people with amd64 and ppc boxes to test07:42
fabbionethere is really nothing i can automate myself there07:42
Keybukzsh: command not found: msgmerge07:49
Keybukanyone !elmo have root on rookery? :)07:52
lamontg'night all08:05
dokomorning lamont08:07
jdubnight lamont 08:08
Keybukgah, I really cannot work this frickin' po file stuff out08:13
Burgundaviacya all08:27
KeybukI really don't think msgmerge is going to work for this stuff08:27
=== Micksa_ [~mslade@203-217-18-166.perm.iinet.net.au] has joined #ubuntu-devel
=== sid77 [~sid77@host98-44.pool8020.interbusiness.it] has joined #ubuntu-devel
sid77hi all08:43
=== Micksa_ is now known as Micksa
=== maswan [maswan@kennedy.acc.umu.se] has joined #ubuntu-devel
=== maswan hopes that a mirror admin isn't unwelcome here
maswanelmo: @ERROR: max connections (25) reached - try again later08:58
maswan(this night)08:58
fabbionepitti: where is USN-4-1 ?08:59
fabbionemdz: ^^09:00
Keybuknot released yet, I'd guess09:08
KeybukI 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 notice09:08
=== Keybuk cackles with pure evil
fabbioneit was on ubuntu-users09:12
Keybukso I've actually "fixed" the .po issue09:12
KeybukI have some Python that strips all the context from a patch and rejiggles the line numbers09:12
=== skyrider [~skyrider@kid.stu.cn.ua] has joined #ubuntu-devel
Keybukso it actually works09:12
Keybukscrew you hippy po comments09:13
Keybuk        if line.startswith(" ") or \                 (line.startswith("-#") and not line.startswith("-#,")) or \                 (line.startswith("+#") and not line.startswith("+#,")):09:16
Keybukthose po file people are as magic-punctuation happy as tom lord!09:16
=== tfheen [~tfheen@111.80-202-211.nextgentel.com] has joined #ubuntu-devel
Keybuk-#, and +#, are a change to the bit of the .po file that says "fuzzy" or "c-comment", etc.09:16
KeybukI got the impression people would be upset with me if I dropped those09:16
fabbionehey tfheen 09:17
fabbionewelcome back09:17
=== tfheen grumbles over vawad being ill
Keybuktfheen: vawad's *always* ill :D09:17
fabbionex.org is taking a long time to build09:17
fabbioneeven with ccache09:17
tfheenKeybuk: she's started to eat networking cards now.09:17
tfheenso, anything fun happened while I was away?09:18
fabbionethe build log for the normal part is at least 2MB bigger than Xfree8609:18
tfheenapart from Warty releasing, obviously.09:18
=== fabbione cries out loud and ask for 4GB of ram
tfheenfabbione: get an amd64 with 4G of RAM and crosscompile? ;)09:24
=== tfheen hides
fabbionetfheen: welcome to send me one :)09:24
tfheenfabbione :)09:24
fabbionemodel name      : Intel(R) Pentium(R) 4 CPU 2.00GHz09:24
fabbioneit's not like i have a slow machine09:24
fabbionewith 1 GB of ram09:24
Keybukdude, it's a P-4 ... that's about 200Mhz in *real* CPU terms <g>09:24
fabbionemodel name      : AMD Athlon(tm) XP 2100+09:24
fabbionei also have this one with 1 GB of RAM09:24
fabbioneit still takes hell of a lot to compile compared to XF8609:24
fabbione7684 -rw-r--r--    1 fabbione fabbione  7852788 Oct 21 12:12 make_world.build.log09:24
fabbione9940 -rw-r--r--    1 fabbione fabbione 10161505 Oct 27 06:56 make_world.build.log09:24
=== fabbione sighs even more
fabbione3020 -rw-r--r--    1 fabbione fabbione  3087544 Oct 21 12:19 make_world_dbg.build.log09:25
fabbione4600 -rw-r--r--    1 fabbione fabbione  4694731 Oct 27 07:24 make_world_dbg.build.log09:25
=== fabbione screams and runs away
tfheenfabbione: you're supposed to _like_ multi-MB build logs.09:27
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
Keybukseems to be working (that's a dropped Debian patch, for obvious reasons)09:29
tfheenKeybuk: you're merging d-i branding?09:30
Keybuktfheen: merging debian and warty to make hoary09:30
fabbionetfheen: oh yeah09:33
fabbioneand for last:09:43
fabbione 868 -rw-r--r--    1 fabbione fabbione   880853 Oct 21 12:22 make_install.log09:43
fabbione1224 -rw-r--r--    1 fabbione fabbione  1245422 Oct 27 07:43 make_install.log09:43
fabbionemore than 50% BIGGER!09:43
fabbioneOH YEAH09:43
Keybukis that X.org, or just daniels? :o)09:46
fabbioneX.org :-)09:47
fabbionei think i need a bigger machine09:47
fabbionei am sure debian buildd will NOT like this09:47
Keybuk* Creating copy of base-files09:48
Keybuk* Considering base-files09:48
Keybuk! 2 patch hunks dropped09:48
Keybuk* Building base-files_3.1ubuntu1.dsc09:48
Keybuk-- cool, most the drops are easy (that's our issue & issue.net change)09:48
=== skyrider [~skyrider@kid.stu.cn.ua] has left #ubuntu-devel []
=== sid77 bye
maswanfabbione: how about an 8-way opteron with 32 gigs of ram?09:57
maswanfabbione: a bit too expensive to get donated though, even if I probably could borrow one for a couple of weeks. :)09:58
fabbionemaswan: useless09:59
fabbioneX build doesn't fork across processors09:59
maswanfabbione: single-cpu speed more relevant?09:59
fabbione32GB would be still good :-)10:00
maswanWell, you could start by fixing that then. :)10:00
fabbionemaswan: send me 2x16GB stick for a Dell Optiplex GX260 :-)10:01
fabbionei won't mind that kind of donation ;)10:01
maswanfabbione: heh. well, sticking within reality, you could get an hp dl585 with 32 gigs of ram too, with only 4 cpus. :)10:02
maswanfabbione: but then, those aren't usually free either10:02
fabbionemaswan: 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
fabbioneKeybuk: "Dear Santa^WMark" ;)10:03
maswanfabbione: see, Keybuk has the general idea of it :)10:03
fabbionethe first step will be to get my bigger office up and running10:04
fabbionesecond step one 19" rack10:04
fabbionethan i will ask Santa to fill it up :P10:04
Keybuknow's the time to decide between air mailing xmas cards, or just e-mailing; isn't it :p10:04
maswanfabbione: a 2.4GHz or so opteron should be among the faster cpus for gcc grunting too, I think.10:05
jameshthe new Athlon 64 4000+'s look nice too10:07
jameshif you just want a single processor machine10:07
maswanwell, yeah, but those only fits a few gigs of ram10:07
maswanof course, they are lots cheaper though10:07
jameshhow much ram do you want?10:07
maswanfabbione thought 32 gigs would be nice :)10:08
fabbionewith 32GB i can mount /usr/src in ram10:08
jameshyou'd need 4 processors for that, right?10:08
fabbioneand compile in it :-)10:08
fabbioneone processor or 2 are more than enough10:08
maswanwell, yeah, I don't know of any board with more than 8 slots/cpu10:08
jameshwhat's the largest DIMM you can get?10:08
=== fabbione does
jameshgiven that each CPU can handle 4 sticks of memory10:09
maswan1GB in practice, 2GB in theory10:09
fabbionemaswan: get a few e10k cpu boards :P10:09
maswanjamesh: I've seen 810:09
maswanfabbione: ah, but those do not have very fast single-cpu performance. :)10:09
jameshmaswan: for a single CPU AMD64?10:09
fabbionemaswan: i know :-)10:09
fabbionethey are still sweet tho10:10
maswanjamesh: well, yeah, you could just run a dl585 with one cpu. :)10:10
jameshmaswan: I've seen 8 slot dual processor AMD64 boards10:10
jameshmaswan: but the second 4 slots could only be used if you plugged in two CPUs10:10
maswanjamesh: well, are those 8 slots in total or 8 slots per cpu?10:10
maswanah, yeah.10:10
jameshsince the memory controllers are on the CPUs10:10
maswanthat's the thing, the dl585 has 8 slots per cpu10:10
maswanhmm.. 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
maswanI think. I haven't looked much at that though.10:11
jamesheither way, you're looking at a lot of money10:13
jameshit might be more economical to buy a bunch of Raptor hard drives and set up a RAID array10:15
=== maswan waves and heads off
=== Ferry [~Ferry@cust.15.118.adsl.cistron.nl] has joined #ubuntu-devel
=== seb128 [~seb128@ANancy-151-1-7-33.w83-194.abo.wanadoo.fr] has joined #ubuntu-devel
=== fabbione cries
seb128that's ok to start working/uploading in hoarty now ?10:42
fabbionedpkg-buildpackage -rfakeroot -uc -us -b10:42
fabbionedpkg-buildpackage: source package is xorg10:42
fabbionedpkg-buildpackage: source version is 6.8.1-0.010:42
fabbionedpkg-buildpackage: source maintainer is Fabio M. Di Nitto <fabbione@fabbione.net>10:42
fabbionedpkg-buildpackage: host architecture is i38610:42
=== pitti [~martin@] has joined #ubuntu-devel
pittimdz: still awake?10:57
seb128hello pitti 10:59
pittiHi seb128!10:59
seb128ok, so nobody knows if we can start working/uploading in hoary if we should better wait a bit ?11:00
pittiseb128: I already uploaded some stuff and was told that it was okay11:02
=== martink [~Martin_Kr@] has joined #ubuntu-devel
seb128yeah, I've seen the mails on hoary-changes, but I was not sure with the syncs11:03
seb128thanks pitti 11:03
=== tuo2 [~foo@adsl-36-114.swiftdsl.com.au] has joined #ubuntu-devel
pittiseb128: I still remember elmo saying that the syncs are finished (the automatic ones, that is)11:03
seb128cool, thanks11:05
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #ubuntu-devel
pittielmo: Morning! Since mdz is already sleeping, can you do the libc unleashing, too?11:15
Keybukbasically there are three sets of things to sync11:16
Keybuk1. things that haven't changed in warty or debian -- easy, nothing to do (so done)11:16
=== Mithrandir [~tfheen@vawad.err.no] has joined #ubuntu-devel
Keybuk2. 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
Keybuk3. things that changed in both debian and warty -- need to merge, I'm working on this at the moment11:17
=== enrico [~enrico@enrico.developer.debian] has joined #ubuntu-devel
enricoHello.  Is someone around that is involved in the Wiki migration?11:33
enricojustdave: around?11:35
=== rburton [~ross@] has joined #ubuntu-devel
=== enrico listens to the sound of one hand clapping
robtaylorpitti: what do reckon to mdz's chroot idea for #2758?11:49
robtaylorpitti: i tried restarting hotplug, and running the same hald test, but didnt seem to change anything11:50
pittirobtaylor: well, I thought hotplug was not the problem11:51
pittirobtaylor: you'd need to run hal not in a chroot, but in the normal file namespace11:51
robtaylorpitti: you think it's kernel level?11:51
pittirobtaylor: I did not extensively deal with the live cd, though11:51
pittirobtaylor: no idea, sorry; please ask amu, he should know much better about the live cd11:52
robtaylorpitti: yeah, it a bit weird, i'm testing now with a more normal usb key and seeing exatly the same symtoms :/11:52
robtaylorpitti: cool. what timezone is amu?11:52
pittirobtaylor: mine, i. e. UTC+211:54
robtaylorpitti: ah cool11:56
=== lifeless [~robertc@dsl-] has joined #ubuntu-devel
=== sivang [~dannyh@] has joined #ubuntu-devel
=== sid77 [~sid77@host98-44.pool8020.interbusiness.it] has joined #ubuntu-devel
=== enrico tries again
enricoWho's involved in the wiki migration?  I'd need to get in contatct with them12:19
seb128why not saying here what you want ? perhaps somebody can help you ?12:20
fabbioneciao enrico 12:20
enricofabbione: ciao!12:20
fabbioneseb128: that sounds familiar :-)12:20
sid77ciao! (i'm italian too ;)12:20
enricoAppearently the pages were migrated12:20
enricosid77: ciao!12:20
enricoHowever, the old wiki has not been locked12:20
enricoI 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 lost12:21
enricoSo, I'd like to ask what is the migration path.  People in the list are puzzled12:21
KamionKeybuk: hm, msgmerge was working absolutely fine for me, seemed to be perfectly automatable in principle12:21
=== migus [~migus@netu.naquadah.org] has joined #ubuntu-devel
enricoThing is, wiki is the way people in ubuntu-doc are making all the work now, and it's been changing without notices for a while12:23
enricoSomeone starts being annoyed of having something different or unclear every time they want to do some work12:23
KamionSteveA 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
fabbionehi sid12:24
=== sid77 is away: Far from here (close to you)
fabbionesid77: please no public away messages12:28
=== __daniel [~daniel@td9091b0b.pool.terralink.de] has joined #ubuntu-devel
fabbionehmm intersting :-)12:40
fabbionepciutils wants to kill ubuntu-base12:40
pittisjoerd: 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-perl12:40
fabbione(bootstrapping hoary)12:40
pittifabbione: I just tried to switch my apt sources to Hoary, but the package files are still empty. Do I sth. wrong?12:41
KeybukKamion: I just couldn't get it to do the right thing, ever12:41
fabbionepitti: they aren't here...12:41
Keybukgiven "warty.po" and "debian.po", I couldn't get one that had the best of both12:41
fabbionepitti: debootstrap warty and then change apt lines12:41
pittifabbione: you mean they are not empty for you? Then it must be a problem of my provider...12:41
KamionKeybuk: really did work fine for me12:41
pittifabbione: my provider has a transparent proxy, maybe this still offers me the empty files12:41
pittifabbione: thanks12:42
KeybukKamion: when I did msgmerge warty.po debian.po; I got the new debian strings, but the warty translations12:42
KamionKeybuk: like I say, you need to produce hoary.po first using debconf-updatepo or whatever12:42
fabbionepitti: i am using a local mirror12:42
KamionKeybuk: I wouldn't do it that way ...12:42
Keybukwhen I did msgmerge debian.po warty.po I got the debian translations, but not the new strnigs12:42
KamionKeybuk: both those approaches are wrong12:42
KeybukKamion: doesn't work for ordinary .po12:42
Kamioncan you leave everything with d-i branding to me then? I don't want to have to redo it all12:42
Keybukyou won't have to :)  I had a lot more luck with mutating the patch12:42
elmoGAR. stupid moin.12:42
=== jdub chuckles.
Kamionordinary .po there's usually some kind of Makefile target to do it12:43
KamionKeybuk: uh12:43
elmodoes anyone know how to get this stupid thing to read it's updated config file?12:43
KeybukI wrote some code to write patches that could apply to .po files, without the # bits getting in the way :p12:43
KamionKeybuk: anything I can review before upload, please?12:43
KeybukKamion: is cooking now12:43
Kamionthose comments are often needed for actual real live translators12:43
KeybukKamion: sure, but you can favour one set over the other12:43
Kamiontrue ...12:44
Keybukit keeps the #, one12:44
Keybukbecause that's important12:44
Kamionthere are others too12:44
Keybukbut #: doesn't even need to be in the patch12:44
Kamion#: should be handled by debconf-updatepo in the debian/po/ case12:44
fabbioneok guys.. good news12:45
=== sid77 is back (gone 00:20:04)
Keybuk*shrug* this way works :)12:45
fabbioneX.org compiles both on AMD64 and ppc12:45
KamionKeybuk: like I say, can I please just have a look at one of them before upload?12:45
fabbionenow it's only question of fixing a few tons of packaging problems12:45
KeybukKamion: dude, someone's going to look at *all* of them before upload12:45
Keybukyou get a source package, debian->hoary diff and "I dropped these" file12:45
KamionKeybuk: you said you were mass-uploading, I believed you :-)12:45
elmojdub: ?12:46
KeybukKamion: yeah, for the ones with no "dropped" I can interdiff debian->hoary and base->warty quickly <g>12:46
Kamion14:37 < Kamion> Keybuk: are you going to be doing a mass upload of merges, or what?12:46
Kamion14:38 < Keybuk> Kamion: I think the basic idea is to do the merges as automatic as possible12:46
Kamion14:38 < Keybuk> for any that it works, I'll upload12:46
Keybuknote "works" :)12:46
jdubelmo: here12:46
elmojdub: do you know how to get moin to reload it's config?12:47
elmoand/or where you involved in setting our moins up?12:47
pittiIs Hoary automatically synced to sid now every day?12:48
=== __daniel [~daniel@td9091b13.pool.terralink.de] has joined #ubuntu-devel
jdubelmo: i set up the original one on rince -> it should just take on the configuration when a new cgi is run12:49
elmopitti: yeah, where it can be12:50
elmojdub: that's what I thought, it doesn't appear to be, tho :/12:50
pittielmo: nice, thanks12:50
=== seb128_ [~seb128@ANancy-151-1-11-54.w83-194.abo.wanadoo.fr] has joined #ubuntu-devel
fabbioneccachetop is cool :-)12:53
enricoelmo: restarting apache?12:53
elmoenrico: done that12:54
KamionKeybuk: fair enough12:55
enricoelmo: still modifiable, though12:55
=== sid77 [~sid77@host98-44.pool8020.interbusiness.it] has joined #ubuntu-devel
KeybukKamion: if there really is some msgmerge magic, that'd be cool -- but I really couldn't get it to produce sane output12:56
Keybukit tended to favour one set of translations over another, completely12:56
KamionKeybuk: note I've done main-menu, cdrom-checker, cdrom-detect, yaboot-installer since I last talked to you12:56
Keybukand dropped others, with no way of actually telling it dropped them without resorting to diff12:56
enricoelmo: sure that the new config is correct?12:56
KamionKeybuk: I used msgmerge successfully on five packages with complex merges12:56
Kamionthe output was exactly what I wanted12:57
sjoerdpitti: no i didn't 12:57
elmoaha.  STUPID moin.12:57
elmoright, it's done12:57
pittisjoerd: it is very likely that it FTBFS for Debian, too12:57
KamionI 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/12:57
KeybukKamion: yeah, and I followed that, and got crap out of the other end12:58
sjoerdpitti: yeah probably.. will check 12:58
enricoelmo: works, cool, thanks!12:58
Keybukentirely commented-out blocks and most of the warty changes dropped12:58
KamionKeybuk: what you told me you tried just above was totally different, though?12:58
KamionKeybuk: entirely commented-out blocks are expected and useful, they act as notes of the unbranded translations12:58
sjoerdpitti: i'm currently checking out pmount, are you planning to create a group in later debian packages or something like that12:58
KamionKeybuk: sounds like you did it the wrong way round12:58
KeybukKamion: I'm applying the debian changes to warty, not the warty changes to debian12:59
KamionKeybuk: me too12:59
pittisjoerd: If Debian agrees to one, it'll be my pleasure :-)12:59
pittisjoerd: plugdev would make Ubuntu and Debian totally compatible...12:59
Kamion11:42 < Keybuk> Kamion: when I did msgmerge warty.po debian.po; I got the new debian strings, but the warty translations12:59
Kamion11:42 < Keybuk> when I did msgmerge debian.po warty.po I got the debian translations, but not the new strnigs12:59
Kamionyou *have* to do the update against new source lines before attempting to msgmerge01:00
Keybukyeah, warty.po there has had debconf-updatepo run over it01:00
Kamionafter patching the rest?01:00
Kamionalso, you need to use a compendium with pre-updatepo-warty.po01:00
Kamionthe compendium is essential01:01
Kamionas far as I can tell01:01
KeybukI couldn't get that to make any difference01:01
Keybukwith/without -C produced the same file01:01
sjoerdpitti: who should agree with that :)01:01
Kamionall I can say is "worked for me" :)01:01
Keybukyeah, I'll beat daf up to find some better way later01:01
KamionI had to do some additional branding, but i expected that01:01
pittisjoerd: well, before introducing a new group, it should at least be discussed on d-devel, or not?01:02
Keybukapplying po-tailored patch hunks seems to work too01:02
sjoerdpitti: for something like that, yeah probably01:02
pittisjoerd: because it has wider use; e. g. libgphoto2 could use it, too01:02
Keybuk(after all, it has the ultimate same effect, "replace this msgstr with this one")01:02
sjoerdpitti: be my guest too propose it then 01:02
sjoerdpitti: libgphoto2 uses camera currently....01:02
pittisjoerd: okay; I can explain the whole mechanism in Ubuntu and ask for opinions01:03
sjoerdpitti: yeah would be nice01:03
pittisjoerd: camera is certainly okay, but does Debian really want camera, usbdrive, pcmciadrive, etc.01:03
pittisjoerd: adding to my TODO list01:03
sjoerdpitti: dunno, i was also wondering about that01:03
=== sid77 [~sid77@host98-44.pool8020.interbusiness.it] has joined #ubuntu-devel
=== T-Bone [~varenet@T-Bone.developer.debian] has joined #ubuntu-devel
Kamionlamont: is somebody sorting out a live CD release announcement? question on ubuntu-devel@ ...01:29
T-Bonelamont: ping?01:30
=== x4m [~max@66-90.242.81.adsl.skynet.be] has joined #ubuntu-devel
Keybuk* Considering evolution01:38
Keybuk! 3 patch hunks dropped01:38
Keybuk* Building evolution_2.0.2-3ubuntu1.dsc01:38
Keybukthat's quite impressive, seeing as it's doing 1.4 -> 2.0 on both sides01:39
lamontKamion: mako said he would put something together yesterday...01:44
T-Bonelamont: i'm facing a lot of "unmet dependencies" errors on stage2...01:45
makoKamion: ergh, yes.. got sucked into cc stuff.. will do01:45
lamontKeybuk: impressive, or scary? :-)01:51
lamontT-Bone: so, um, build them.. :-)01:51
lamontT-Bone: example missing build-dep?01:52
T-Bonelamont: heh, i'm scared that if they are missing, it's because they didn't build in the first place ;)01:52
T-Bonelamont: i have also stuff much more scary:01:52
T-Bonecpp-3.3: installed (negative dependency)(but version ok 1:3.3.4-9ubuntu5 << 1:3.3.3-0pre1)01:52
T-Bonelook at "<<"01:53
T-Bonethat's the build log for xfree86. You can take a look at it, it's rather small (3k)01:54
lamontT-Bone: url?01:54
lamontor hostname?01:54
Keybuklamont: well, it means either my hunk-at-a-time patcher works ... or there's a major bug in it noticing patch failures :p01:55
lamontKeybuk: yeah.  cool.01:55
T-Bonelamont: http://envy.esiee.fr/~varenet/logs/xfree86_4.3.0.dfsg.1-6ubuntu25_20041027-124801:55
lamontt-bone: the neg dep is fine01:56
lamontit's the lack of libglide3-dev that's fatal.01:56
T-Bonebut isn't it scary that it says "1:3.3.4 << 1:3.3.3" ?01:57
lamontit's a build-conflicts01:57
Keybuklamont: what's cool is that it spins rookery's pid counter about 6 times on a run <g>01:57
lamontit 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
lamontKeybuk: lol01:57
lamontT-Bone: how did the stage1 glide build go?01:58
Keybukthat's an sbuild message, isn't it?01:58
thomKeybuk: you're a bad man01:58
Keybukthom: I am?01:58
T-Bonelamont: do you know the package name? I can't find a stage1 log for "*glide*" :-/01:59
lamontt-bone: apt-cache show libglide3-dev | grep ource01:59
lamontSource: glide01:59
T-Bonelamont: then i assume it didn't go01:59
T-Bone[varenet@envy ~] $ grep glide build_list01:59
T-Bone[varenet@envy ~] $ 01:59
T-Bonelooks like it wasn't kept by quinn-diff02:00
lamonttry explicitly building it against stage1?02:00
T-Bonewill do02:01
=== nmf [~nmf@] has joined #ubuntu-devel
pittilamont: Hi, howdy!02:03
pittilamont: Has the glibc issue been resolved?02:04
fabbionehey T-Bone 02:04
lamontpitti: yeah.02:04
pittilamont: mdz is sleeping, can you unleash the package to the security archive, too? (and call amber)02:05
lamontpitti: in debian/control.in/libc6, g/Architecture:/s/:/: amd64/02:05
=== lamont can't, but I bet elmo can.
=== lamont wonders how mark is doing on getting those elmo-clones in.
T-Bonefabbione: howdy!02:06
pittilamont: now it builds also on amd64? Must I upload a package with this change or did you already do this?02:06
lamontubuntu2.2 is in the archive02:06
lamontwell, source and binary are sitting there waiting to enter the archive when amber gets goosed02:06
=== jamesh [~james@203-59-217-232.dyn.iinet.net.au] has joined #ubuntu-devel
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #ubuntu-devel
KamionU := $(CURDIR)/debian/udev-udeb02:07
KamionI just KNOW this is going to confuse me02:08
pittilamont: argh, I have to wait for mdz anyway because only he can actually approve the mail to u-s-a02:08
lamonthoary.all.amd64:Total 1130 package(s) in state Needs-Build.02:08
lamonthoary.all.i386:Total 472 package(s) in state Needs-Build.02:08
lamonthoary.all.powerpc:Total 674 package(s) in state Needs-Build.02:08
Kamionwhat's stuck with amd64?02:08
lamontKamion: me.  only 2 buildds atm, while the other 2 both have 302:09
lamontthat and I had one of the 2 offline for a bit over an hour last night02:09
lamontKamion: I need to resurrect king now that it's happy again.02:11
fabbioneT-Bone: how is going the ia64 port?02:12
T-Bonefabbione: it's going ;^)02:12
T-Bonei'm experiencing build failures that have to be solved on a case-by-case basis, so that takes time02:13
T-Bonefabbione: i'm currently progressing in "stage2", aka "warty built against warty"02:13
Kamionhm ... I bet d-i doesn't have /bin/mountpoint02:14
fabbioneT-Bone: cool. please let me know when you have a working pure warty chroot :-)02:14
T-Bonefabbione: heh. Don't worry, when that happens, i'll let people know!02:15
=== x4m [~max@228.190-201-80.adsl.skynet.be] has joined #ubuntu-devel
=== __daniel [~daniel@td9091b13.pool.terralink.de] has left #ubuntu-devel []
thomlamont: what's the deal with gnutls?02:45
lamontthom: ftbfs?02:45
lamontdoesn't seem to be...02:45
thomlamont: no, no. gnutls11 v gnutls1002:46
=== sivang [~sivan@box79162.elkhouse.de] has joined #ubuntu-devel
sivangmorning folks02:46
thomare we standardising on 11 for hoary?02:46
lamontoh that... /me needs to request a sync of gnutls11 from debian (new package and all that...)02:47
elmothom: yes02:52
elmolamont: no, you don't02:52
elmoit's there02:52
lamontthom: metric boatload of changes if we don't, and no reason not to...02:52
lamontelmo: even better.02:52
elmopeople just need to start merging packages that are being held back02:52
robtayloramu: alive yet?02:52
=== lamont didn't see it..
elmolike libldap202:52
robtaylorlamont: hal doesnt seem to work on the livecd...02:57
fabbionedaniels: please nuke Xprint out of X.org02:59
fabbionedaniels: even upstream02:59
fabbionethey don't deserve to live02:59
fabbioneTHEY MUST DIE NOW03:00
Micksayou don't like them?03:02
thomxprint is the stupidest idea in the world. EVAH.03:03
Micksawhat's wrong with the idea?03:04
=== x4m [~max@105-86.242.81.adsl.skynet.be] has joined #ubuntu-devel
danielsfabbione: i'm working on killing it upstream, so is everyone else03:06
fabbionedaniels: thanks03:06
danielsfabbione: remember how I asked you not to ship it? :)03:06
fabbionedaniels: 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
lamontrobtaylor: wonder if it's just that hal doesn't get started...03:07
danielsfabbione: xprint is total bong03:08
lamonttry sudo /etc/init.d/hal start (or whatever it is...)03:08
=== enrico bans the xprint people from earth
=== fabbione opens the "black bag" with the european nuclear sodomotron control
lamontenrico: does that mean they'll be the first to get to mars?03:08
enricolamont: probably.  So they are the one that have to build the toilets03:09
robtaylorlamont: nope worse than that... see http://bugzilla.ubuntu.com/show_bug.cgi?id=275803:09
robtaylorlamont: basically there seems to be a fundamental disconnect somwhere between the kernel and hald03:09
lamontrobtaylor: like that livecd uses a completely different hw detection mechanism?  would that explain it?03:10
fabbionelamont: no... they won't be the first anyway03:10
fabbionelamont: http://www.fabbione.net/bla/marte.jpg03:10
fabbionemarte = mars03:10
lamontthe mars expedition folks here in the states are headquartered in Boulder, CO.03:11
lamontfabbione: cool pic03:11
Mithrandirfabbione: sheez, put ut some decent pizza place at least. ;)03:11
robtaylorlamont: 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
lamontrobtaylor: could be.  is dbus et al even running on a booted livecd?03:12
robtaylorlamont: you mentioned that mini_fo has problems with fs mapped usix sockets? maybe its this?03:12
lamonttcp sockets03:12
lamontspecifically, polling on same with select03:12
lamont==> apt http repositories fail03:13
robtaylorlamont: yep, dbus is running03:13
robtaylorlamont: oh, thats very odd03:13
=== robtaylor wonders why something in VFS land would effect berkeley sockets...
=== Kamion builds a test d-i image with hotplug and udev. Wonder how well it'll work ...
Kamionoh, shit, I need to be gone. Back in a few hours.03:14
lamontrobtaylor: because dereferencing a null pointer in the fs layer before you _get_ to berkeley sockets is still fatal03:14
robtaylorKamion: laters03:14
robtaylorlamont: sounds pretty scary, one way or another =)03:15
lamontyeah - still working out the right anagram expansion for minifo03:15
lamont(NFS == Not a File System"03:15
robtaylorMight interact nastily in filesystem operations?03:16
robtaylori think we got it ;)03:17
lamontheh.  that's the clean version03:17
justdaveenrico: pong03:17
enricojustdave: hey! :)03:19
enricojustdave: solved, thanks03:19
robtaylorlamont: it just seems that the kernel isnt running hotplug ..03:20
robtaylorlamont: if i make hotplug.fuctions write stuff to a file whenever its invoked, nothing ever gets written to that file..03:21
robtaylorunless somewhow the kernel is invoking hotplug in a different root ... is that even possible?03:22
lamontthere's the real root, the basemod root, and the mainmod root.03:30
lamontpick one. :-(03:30
robtaylorbut 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
lamonthence mdz's comment about chroot vs pivotroot03:32
robtaylorhmm. 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
=== lamont decides to go catch a nap, or finish last night's sleep, or whatever you want to call it...
robtaylorok, laters lamont03:33
lamontthis up at 0530 crap is really annoying this night person. :-)03:33
lamontback in an hour or two03:33
robtaylorheh. slerep tight03:33
fabbionelamont: are you trying to turn into me, waking up at 5am each day?03:34
lamontfabbione: 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.03:35
fabbioneoh yeah...03:36
fabbioneforgot about the kids going to school03:36
lamontotoh, my split up day does make it easier for me to catch the non-US crowd(s).03:36
lamontanyway,  bbl03:37
=== [Clint] [~c123p456@user-12hdtek.cable.mindspring.com] has joined #ubuntu-devel
=== cc [~byte@byte.fedora] has joined #ubuntu-devel
=== robtaylor_ [~robtaylor@] has joined #ubuntu-devel
=== robtaylor_ is now known as robtaylor
thom*sigh*; firefox has 24 bugs04:27
Mithrandironly 24?04:29
thomMithrandir: most of them are upstream or weird crashers with flash or java04:29
=== sid77 [~sid77@host98-44.pool8020.interbusiness.it] has joined #ubuntu-devel
Mithrandirprobably "if you feed it this invalid bytecode, the JVM will do funny things and crash FF"04:31
=== irc [~stephane@eratostene.fi.infn.it] has joined #ubuntu-devel
=== fabbione kills XPrint from the Imakefiles EVERYWHERE
=== Mithrandir gives fabbione some beer
fabbioneMithrandir: thanks :-)05:05
fabbioneroot access at fd.org is enough05:05
Mithrandirwe could just sit on daniels next time we meet him, until he gives us?05:05
fabbionerm -rf /var/lib/cvs/xorg/xc/programs/Xserver/XpConfig would be enough05:05
fabbioneMithrandir: eheheh05:06
=== daniels [~daniels@202-44-183-17.nexnet.net.au] has joined #ubuntu-devel
thomok. firefox in warty is 0.99+1.0PR.1+revertedto0.9.3-0ubuntu305:13
=== daniels stares at thom.
elmoseb128: ?05:13
seb128elmo: ?05:13
thomis 1.0~0.10.1-0ubuntu1 a reasonable version for hoary? (can we use ~ yet?)05:13
elmoseb128: you're only doing ubuntu uploads were the are ubuntu changes, right ?05:14
mdzgah, did pitti leave?05:14
elmothom: 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 katie05:14
seb128elmo: hum ? I've uploaded new versions that are not and debian, and merged warty changes for the *gksu*05:15
danielsthom: hmm, since really < reverted, you could have 0.99+1.0PR.1+seriously+1.0PR.1-0ubuntu105:15
elmoseb128: ok, I couldn't tell from your changelog, if there still warty changes or not.. just wanted to double check05:15
seb128elmo: ok, that's fine, thanks for checking :)05:15
=== thom smites daniels with great and terrible wrath
danielsor 'totally+1.0PR.1'05:16
thomi really don't want to go down the 0.999 road05:17
sid770.999 is evil (almost)05:18
danielssid77: as opposed to 0.99+1.0PR.1+seriously+1.0PR.1?05:19
ajelmo: does dak logging/etc deal with ~ okay now?05:19
sid77daniels, that is evil (completely)05:20
thomso the quick answer was "no, you can't use ~. KTHXBYE"05:20
elmoaj: no, that was implicit in "unblock", i.e. use something other than tilde-seperated-value for logging :)05:20
ajelmo: heh05:21
elmothom: no, the katie thing can and should be fixed - the answer to your question is independent to that05:21
ajthom: "1.0~0.10.1-0ubuntu1" is hella complicated; the 0.10.1 can't be just pre1?05:22
thomaj: probably, yes05:24
=== Keybuk [scott@descent.netsplit.com] has joined #ubuntu-devel
thommdz: ^ ?05:26
mdzthom: ?05:27
thommdz: thoughts on firefox versioning?05:27
mdzthom: for going back to 1.0PR?05:27
Keybuk1.0~PR ?05:27
thom(with appropriate patches)05:27
thomKeybuk: my suggestion was close to that, yes05:28
Keybukjdub: sorts less than null, so 1.0~ < 1.005:28
Mithrandirjdub: tilde mean "less than" in newage debian versioning speak.05:28
=== Mithrandir ducks
Mithrandiror less than zero05:29
Mithrandirminus epsilon05:29
Mithrandiror something05:29
Keybukit's a bugger to explain, isn't it <g>05:29
aj"actually, immediately prior to what i just said" kinda works05:29
aj"it's version 1.0 -- actually just before 1.0 -- sub version pre1"05:30
=== Keybuk calls them pre-versions
jdubperhaps "... but i'm bullshitting you" works05:30
ajnah, better to reserve that for the 0.999-but-really-1.0pre1 versions out there05:31
mdzthom: '~' gives me the fright05:32
Keybukmdz: any particular reason?05:33
ajmdz: ooo, costume ideas for halloween05:33
mdzKeybuk: because nobody has used it for anything particularly important yet05:33
Keybukthat's because the Debian archive won't accept them until dpkg 1.10 is in a stable release05:33
mdzbut of course, now is the time to break it if it 's going to break...05:33
=== Mithrandir tries to imagine aj dressed up as a tilde.
mdzit needs an elmo signoff05:33
KeybukI was kinda nervous, then I read the code and it was a "woah! that's actually really elegant" when I saw how it was implemented05:34
aj/msg elmo quick: type /quit ~ is okay by me!05:34
ajKeybuk: (of /course/ it was really elegant...)05:34
fabbionehey mdz05:35
aj(i mean, really -- code getting into dpkg without being elegant? the very thought!)05:35
Keybukaj: yeah, I mean, it's so great that I could merge the bzip2 stuff in and being totally sure it would just work05:35
Keybukthe fact it didn't even *compile* because someone typo'd the #ifdef so never actually ever ran it was a figment of my imagination05:36
ajKeybuk: i guess you take drugs for that now though?05:36
Keybukaj: I must be, for agreeing to maintain the fucking thing :p05:37
MithrandirKeybuk: did you ever agree to it?05:37
ajKeybuk: oh dear, never a good sign when you don't know what drugs you're taking05:37
MithrandirI thought we just made you?05:37
elmomdz: I'll need to unbreak katie - but I thought you had concerns about back compat or so?05:38
mdzelmo: depends on whether sarge releases before hoary05:39
mdzmozilla-firefox isn't even in woody, is it?05:39
Keybukit didn't *exist* when woody was released05:39
jdub(haw haw)05:40
aji don't see how it matters even if hoary releases before sarge?05:40
mdzaj: if hoary releases before sarge, we will support upgrades from woody to hoary05:40
ajyou've got 1.0.1 in woody, 1.1 in warty, 1.2~1 in hoary, 1.2 in sarge05:40
ajyour 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/apt05:41
Keybukmdz: but we don't support upgrades from woody to warty ?05:41
mdzKeybuk: of course we do05:43
Micksalifeless: I'm still undecided on whether to come on on friday.05:44
mdzthom: why don't we stick with what Debian is doing?  they seem to have 0.10.1+1.0PR05:45
mdzthat's newer than 0.99+1.0PR.1+revertedto0.9.3-0ubuntu305:45
=== Keybuk covers his eyes!
Mithrandir0.100.1+1.0PR ?05:45
Keybukit really isn't dude05:45
Keybuk0.10 < 0.9905:45
mdzoh, missed the 0.9905:46
thom0.10.1 so isn't hiugher05:46
mdzwtf is 0.99?05:46
mdzI read it as 0.905:46
Keybukmdz: nearly but not quite05:46
sid77what about a 0.9.3-<current version>?05:46
Keybuksid77: 9 < 9905:46
thom0.99 was my screwup early on05:46
mdzthom: if elmo is ok with ~, I'm ok with ~05:46
mdzelmo: you need to unbreak katie, or you need to remove the fascist check that forbids ~?05:47
mdzI thought katie used apt's version comparison code05:47
sid77Keybuk, right, but doesn't it start from 0.9.3?05:47
elmomdz: it does - I thought apt supported ~ ?05:47
mdzelmo: it does05:47
mdzelmo: so what unbreaking is required?05:48
thommdz: 16:20 < elmo> aj: no, that was implicit in "unblock", i.e. use something other than tilde-seperated-value for logging :)05:48
Keybuksid77: 0.9 is less than 0.99 because 9 is less than 9905:48
elmomdz: I use ~ as a separator in log files.   go me.05:48
mdzelmo: oh, rad05:48
elmoand also there's a regex which checks for "validity" which needs updated, but that's trivial05:48
thomyay for intra debian communication05:48
=== Keybuk looks at that "*please* implement Enhances" bug
elmothom: 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 versions05:49
elmoerr, previous work, obviously05:49
Keybukwe're fast running out of silly characters to use for things05:49
sparkesyay, utf05:49
sid77Keybuk, maybe we should introduce a new number space where 99 < 9 :)05:50
=== lupus_ [~lupus@kn-ivl-2.kuleuven.net] has joined #ubuntu-devel
Keybukthom: does firefox use libtool anywhere?05:53
lifelesssid77: aleph-1 ?05:53
lifeless(as opposed to aleph1)05:53
thomKeybuk: oh christ05:54
sid77lifeless, geeee...05:54
mdzelmo: 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?05:54
elmomdz: I can probably fix it tonight - shouldn't take long05:55
mdzKeybuk: will libtool require un-breaking as well?05:55
=== Keybuk considers what other silly character to use as $IFS <g>
mdzKeybuk: how bad is it?05:56
elmowhy would libtool care about the Debian version ?05:56
Keybukmdz: you know Libtool just flatly falls on its face and flops about when you have spaces in file or directory names?05:56
thomelmo: the ~ is in the upstream version?05:56
Keybukelmo: mozilla-firefox-1.0~1/debian/...05:57
elmothis is going in the filename??05:57
Keybukelmo: goes in the directory name, doesn't it05:57
=== elmo runs away screaming
mdzKeybuk: sure, but we're not giving it spacess, we're giving it tildes :-)05:57
thomok, so maybe i'll work on some other stuff for a while :-)05:57
Keybukmdz: libtool uses IFS="~" internally for some things05:58
Keybuk(multi-command link lines)05:58
mdzit's starting to sound like it would be easier to fix dpkg and apt :-P05:58
seb128jdub: 05:58
seb128<CIA-1> cneumair * gnome-panel/gnome-panel/ (10 files): (log message trimmed)05:58
seb128<CIA-1> Fix bug #143963 [http://bugzilla.gnome.org/show_bug.cgi?id=143963] : "Remove Distro Menu".05:58
seb128jdub: finally :p05:58
jdubhooray! :)05:58
=== robtaylor [~robtaylor@] has joined #ubuntu-devel
Keybukthom: http://people.ubuntu.com/~scott/keybuk-tilde-beep.patch is the libtool patch06:06
danielsKeybuk: dude, BEL?  seek professional help.  like, three weeks ago.06:07
Keybukdaniels: well, it's unlikely to appear in filenames <g>06:07
Keybuk(well, unless your cat walks across your keyboard while in vi)06:08
danielslibtool has destroyed your brain06:08
thomholy mother of god save us.06:09
Keybukit was ^M, but some editors "helpfully" strip that06:09
=== daniels stares at Keybuk.
ajKeybuk: 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
Keybukaj: that's the other solution, yes.  fix dpkg-source06:10
Keybukthat's actually less sick than it sounds, because we already don't put the epoch and ':' in filenames06:11
ajapt does, but url escapes them06:11
Keybuktar gets somewhat upset when asked to do 'tar czf blah_1:1.0.orig.tar.gz blah-1:1.0'06:11
Keybukit thinks blah-1 is a hostname to rcp 1.0 from06:12
Keybukyou've never encountered this before?! :p06:12
aji don't use colons in filenames :)06:12
ajnor do i use rsh :)06:12
Keybukthis is the reason dpkg hides the epoch from the filenames06:12
Keybukit breaks things :p06:12
thomthat sounds like a much saner approach for ~ too06:14
ajwow, tar sucks06:14
elmoKeybuk: WTF dude, neither had you till a week ago - I told you about it in London :-P06:14
Keybukelmo: true :o)06:14
ajgeez, elmo's such a spoil sport06:14
Keybukthat was the first time I'd encountered it <g>06:14
Keybukand, admittedly, probably the first time I'd ever tried to use ':' in a filename06:14
=== irc [~stephane@eratostene.fi.infn.it] has left #ubuntu-devel []
ajtar czvf - foo:bar > foo:bar.tgz works though :-/06:16
ajahh, gnu tools. gotta love 'em.06:16
=== Keybuk ponders which of s/~// and s/~/%7f/ is the least sick
robtaylorKeybuk: s/~/LESSTHANONOTREALLYMAYBE/ ;~)06:19
robtaylorlets keep that clear semantic information intact ;-) 06:19
jdubnight all06:22
robtaylornight jdub 06:24
=== sid77 bye
Keybuk%7F is ugly06:25
=== Keybuk squirms
Keybukdpkg-source.pl: extracting build-essential in build-essential-11%7F06:25
Keybukmuch cuter without it06:26
thomthat is quite nasty06:26
Keybukdpkg-source.pl: extracting build-essential in build-essential-11_06:27
Keybuka little nicer06:27
robtaylorheh. more mathematicallys s/~/)/06:28
robtaylor(out of interest.. whats the problem with just using ~ in the filename?)06:30
Keybukrobtaylor: libtool go boom06:32
robtaylorKeybuk: ah . boh06:32
=== bob2 [rob@bob2.user] has joined #ubuntu-devel
Keybukthough I actually can't make it fail now, amusingly06:34
robtaylorKeybuk: is there a particular reason it might choke on ~?06:34
Keybukrobtaylor: yeah, uses IFS=~ internally06:34
Keybukthough it's now working on places it should fail06:35
Keybukar cru .libs/libtest.a  /home/scott/tmp/tilde~test/test2.o06:36
Keybukranlib .libs/libtest.a06:36
Keybuk-- 06:36
Keybukthat should fail, theoretically06:36
Keybukmaybe it's just 1.4 which breaks, in which case I don't care06:38
Keybuk(you can see I really tested this, can't you :p)06:38
robtaylorisnt that waht users are for? ;)06:38
Keybukthom: can you try building firefox in a dir with ~ in it ... I have a hunch it'll work06:40
Keybukmy libtool test suite *cough*gnome*cough* builds ok06:41
Keybuk1.4 does fail06:41
Keybukbut 1.5 actually doesn't eval the variables until inside the loop (at which point the IFS=~ doesn't matter)06:42
Keybuk    save_ifs="$IFS"; IFS='~'06:43
Keybuk    for cmd in $cmds; do06:43
Keybuk      IFS="$save_ifs"06:43
Keybuk      eval cmd=\"$cmd\"06:43
Keybukso the variable with the directory name in it doesn't get expanded inside $cmd until afterwards06:44
bluefoxicyWhere 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?06:45
Keybukgah, pretty initscripts make merging hurt06:56
Keybukzsh: segmentation fault (core dumped)  ls -bFh --color=auto -d ../sources/*/results06:59
Keybukgee... thanks06:59
thomKeybuk: it's still running, but it's not blown up07:03
=== Kamion returns
Keybukelmo: do you want a list of things we can actually just sync from Debian again?07:04
Keybukmost of lamont's libtiff4-fest for example07:05
Kamionhm, I do like the way my random test Ubuntu installs are set up with a random language07:06
Kamionand keyboard layout; makes for extra fun when typing passwords07:06
=== robtaylor is now known as robtaylor|away
=== enrico [~enrico@enrico.developer.debian] has joined #ubuntu-devel
Kamionhooray, my udev-ised test CD ... totally blows up07:12
thomrock on07:13
mdzKeybuk: do you have a merged alsa-utils?07:14
Keybukyup, want it?07:15
thom < Tybstar> it would be cool if CUPS had a reasonable default paper size depending on the location you picked at install07:15
mdzthom: yes, it would07:16
thom(he works for specifix, who do conary)07:16
thomoh dear:07:17
thom18:15 < sri> fer: just think how awsome the porn would be, various actions related to each stage woo..07:17
thom18:15 < fer> sri: wops07:17
thom(on graphical boot)07:17
=== bluefoxicy pokes people
thom18:16  * sri can imagine hww delicious fscking would be :)07:18
thom18:16 < fer> linux-2.6.9-p0rn107:18
thom18:16 < fer> and mount!07:18
Keybukmdz: http://people.ubuntu.com/~scott/alsa-utils/07:18
Keybukthere's an oddness there, it looks like Debian dropped 90_debian_alsaconf.dpatch07:19
Kamionah, busybox mount doesn't support --bind, hmm07:19
Keybukand there's a few Debian changes to control in .dropped that didn't apply07:19
mdzthom: CRACK07:23
thommdz: *g*07:23
mdzKeybuk: if debian dropped that patch, why is it still in debian/patches in the output?07:23
Keybukmdz: because we changed it07:23
mdzwe what?07:23
Keybukwe changed 90_debian_alsaconf.patch07:23
Keybuk-+  cfgfile="/etc/alsa/modutils/1.0"07:24
Keybuk++  cfgfile="/etc/modutils/alsa-base"07:24
mdzok, that's not important, since we later dropped alsaconf entirely07:24
Keybukbit 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
mdzI wonder if it's fixed in 1.0.607:24
Keybuk_warty.patch is the base->warty patch, _debian.patch is base->debian, _hoary.patch is debian->merge07:25
Keybukso you pretty much side-by-side compare _warty.patch and _hoary.patch to see whether anything went wrong07:26
mdzwhere did all these control file changes come from?07:27
mdzthey're not in the changelog07:27
=== mdz looks at fabbione
Keybuk"fabbione" :)07:27
Keybukthe ones in .dropped are *Debian* changes07:27
Kamionmmmkay, I wonder if d-i trunk is hosed at the moment07:27
Keybukthe ones in _warty.patch were our changes07:28
mdzwhere is .dropped?07:28
Kamionthere's some very suspicious pissing about with /proc/self/fd/0 and /dev/console07:28
Keybuksame directory07:28
mdzah, my 'get *' didn't get it07:28
mdzoh, yes it did07:28
Keybukthat basically logs the IKEA "I put it together, what's this bit for?" hunks07:28
mdzI thought you meant literally ".dropped"07:28
bluefoxicyNoo ne?07:30
mdzKeybuk: the reason that one fell out is because fabio changed the Maintainer: field in the ubuntu branch for some reason07:30
Keybuknot sure why the description fell out, I suspect just because it was a bit too close07:31
thombluefoxicy: the ubuntu devel mailing list would seem to be appropriate07:31
mdzKeybuk: resolved, uploaded07:34
Keybukcool, I'm just finishing off the last of the packages without a .dropped (basically checking things are sane)07:34
mdzKeybuk: did you get official word from elmo about your key?07:34
Keybukthen I'll stick them all somewhere07:34
Keybukmdz: not yet07:34
mdzmdz@jackass:~ $ gpg --keyring /srv/keyring.no-name-yet.com/keyrings/ubuntu-keyring.gpg --list-keys remnant07:35
mdzpub  1024D/84AD676C 1999-12-30 Scott James Remnant <scott@netsplit.com>07:35
mdzlooks to me like it's there, as it should be07:36
Keybukthat's the wrong one07:36
bluefoxicythom:  oy now I have to find those-07:36
KeybukC978C8AE  is my canonical key07:36
=== bluefoxicy is subscribed to something like 36 MLs
bluefoxicythom:  no better suggestions?07:36
mdzKeybuk: does there exist a copy of that key with useful signatures on it?07:40
mdzI seem to have one in my keyring, but it has zero signatures07:41
seb128hum. Any way to get a sync for librsvg2 2.8.1-1 ? The warty changes are in the new debian version too 07:41
mdzseb128: sure, just email elmo07:41
Keybukmdz: should be signed by daniels, fabbione, lalo and debonzi07:42
mdzKeybuk: but where?07:42
mdzit's not on pgp.net or keyring.d.o07:42
mdzKeybuk: if you only keep the signatures locally, it doesn't count :-)07:42
Keybuksubkeys.pgp.net has them07:44
seb128elmo: any way to get a sync from debian for librsvg2 2.8.1-1 ? The warty changes are in the new debian version07:47
elmoseb128: done07:48
seb128thanks !07:48
bob2 hm, why is "apache-common" in main, while the rest of apache 1 is in universe?07:50
elmoit's not07:51
elmoapache-dev is in main too07:51
elmofor php407:51
bob2hrm, ok07:51
bob2well, the apache 1 daemon is in universe at least07:51
bob2oh, ew, some of apache 1 is in main so libapache2-mod-php4 can build?07:52
Kamionhm, current grub seems to build-dep on type-handling too ...07:53
mdzthat type-handing crack needs to go07:54
mdzgrub.maintainer == type-handling.maintainer, I think07:55
=== |trey| [~trey@ip68-230-75-109.ph.ph.cox.net] has joined #ubuntu-devel
=== |trey| [~trey@ip68-230-75-109.ph.ph.cox.net] has joined #ubuntu-devel
Kamionmdz: yes08:02
Kamionmdz: you misspelt "dealer"08:02
=== mdz nods
=== __daniel [~daniel@td9091b13.pool.terralink.de] has joined #ubuntu-devel
mdzlamont: ping?08:03
mdzKamion: are you running britney or anything like that over hoary yet?08:03
mdzdo we have some idea what the dependency breakage is like?08:03
Kamionmdz: elmo runs it now, I just mirror it08:04
Keybuk-Maintainer: Tommi Virtanen <tv@debian.org>08:04
Keybuk+Maintainer: Fabio M. Di Nitto <fabbione@fabbione.net>08:04
elmomdz: yeah, it's bad, 'cos of gnutls mostly AFAICS from a quick glance08:04
elmowhich should be building now08:04
Kamionmdz: but it's in the same place, yes08:04
mdzKeybuk: which package is that?08:04
mdzKamion: found it, thanks08:05
Kamion__daniel: script that manages Debian's testing distribution; we run it in a very cut-down form to check for dependency breakage08:05
mdzelmo: gnutls caused binary deps to break?  or build-deps?08:05
elmomdz: binary, gnutls10 -> 11 transition08:05
Keybukmdz: uh, python-osd I think08:06
elmoe.g. libldap2 is uninstallable08:06
KeybukI went past it pretty quickly with a giggle08:06
__danielKamion: wow - that sounds cool - every time i hear about debian / ubuntu infrastructure i'm absolutely impressed08:06
mdzah, right, all the libgnutls10 deps on our existing binaries08:06
mdzthat sort of sucks08:06
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 worked08:07
Kamion__daniel: I got bored of doing that very quickly :)08:08
__danielkamion: 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
__danieli'm still very perplexed by it all :-)08:09
=== nmf [~nmf@] has joined #ubuntu-devel
=== stevedeo [~steve@static24-72-62-185.regina.accesscomm.ca] has joined #ubuntu-devel
mdzelmo: so once gnutls is built, we need to rebuild pretty much everything?08:28
=== maskie [~maskie@196-30-111-93.uudial.uunet.co.za] has joined #ubuntu-devel
=== |trey| [~trey@ip68-230-75-109.ph.ph.cox.net] has joined #ubuntu-devel
lamontmdz: ack08:38
mdzlamont: what does the hoary buildd picture look like?08:39
=== x4m [~max@85-117.240.81.adsl.skynet.be] has joined #ubuntu-devel
lamonthoary.all.amd64:Total 894 package(s) in state Needs-Build.08:40
lamonthoary.all.i386:Total 1 package(s) in state Needs-Build.08:40
lamonthoary.all.powerpc:Total 181 package(s) in state Needs-Build.08:40
lamontworking on that amd64 thing now08:40
lamontwhich is part of why it's behind - took one buildd offline to clone the 4 chroots08:41
mdzlamont: thanks08:42
mdzlamont: also, how many packages will we need to rebuild to fix the gnutls situation?08:43
lamontgnutls11 is in the archive, with stuff depwaitng on it - you're after that count?08:44
lamonthoary.all.amd64 1808:44
lamonthoary.all.i386 2108:44
lamonthoary.all.powerpc 2008:44
lamontpackages dep-wait libgnutls11-dev08:45
mdzlamont: I mean, we have a bunch of pakcages which depend: libgnutls1008:49
mdzdo all of those have new versions which build-dep on the new gnutls?08:49
mdzor will we need to build some of them explicitly?08:49
lamontah, will check.08:50
lamontfwiw, gnutls10 is there, but I gather we want it to go away.08:51
lamont70+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
lamontmeanwhile, fixing amd6408:52
mdzlamont: gnutls10 binaries will vanish once gnutls11 binaries appear, no?08:54
mdzlamont: from the numbers you gave above, it looks like i386 dust is quite settled already08:55
elmomdz: no08:56
elmoit's two separate source packages08:56
T-Bonelamont: i'm currently under 50% packages successfully built on stage208:56
mdzso we don't actually have a problem, then, do we?08:56
mdzthe stuff which depends on gnutls10 will be satisfied08:57
elmomdz: e.g. libldap2 is uninstallable08:57
elmobecause although both are there, they aren't co-installable08:58
elmoso we need to get our libraries transitioned over ASAP08:58
mdzI thought Debian did this transition for sarge08:58
mdzis it blocked by packages which need merging?08:59
Kamionthere were a few transitions involved, I believe08:59
mdzah, openldap2 is branched08:59
elmoI mentioned this last night - probably should filed a bug, sorry09:00
mdzKeybuk: openldap2?09:00
elmobtw - 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?09:00
Kamionanyone know what's responsible for starting udevd?09:01
KamionI'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 it09:01
lamontelmo: python? :-)09:02
mdzopenldap2 -2ubuntu1 and -2ubuntu2 seem to be present in Debian09:02
Kamioncomponent overrides? :-)09:02
lamontsounds like the component tuple needs to grow by suite-name...09:02
lamontmdz: do I want to know who the uploader was?09:02
elmoremember the pool is split by component09:03
mdzlamont: no, I mean the changes09:03
mdz-2ubuntu1 and -2ubuntu2 were ubuutu backports from Debian09:03
Kamionelmo: oh, toss09:03
elmoso 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/universe09:03
Kamionelmo: hardlinks?09:03
elmooh, hmm, then again.. why does it? :>09:03
mdzKamion: weird, re: udevd09:03
Kamionelmo: mirrors would probably kind of like it to09:04
mdzKamion: I didn't think anything other than S04udev started udevd09:04
Kamionat least people who don't mirror pool/universe/ (in the vice-versa situation)09:04
lamontmdz: ah, that helps09:04
Kamionelmo: such as, hm, just for example, cdimage :-)09:04
Keybukmdz: yeah, there's a openldap2 -- am going to re-run these though with the new .po handling Kamion's just helped me with09:04
mdzKeybuk: openldap2 doesn't have any .po file changes09:04
=== nasdaq4088 [sdfsd@tkp-ip-nas-1-p232.telkom-ipnet.co.za] has joined #ubuntu-devel
Keybukok, hang on then09:05
Kamionmdz: yeah, I'm stumped09:05
mdzKamion: maybe the udev hotplug hook starts udevd if it isn't running?09:06
=== madduck [~madduck@madduck.developer.debian] has joined #ubuntu-devel
mdzKeybuk: fetching, thanks09:06
Keybuknote that the stuff in .dropped *isn't* identically present in Debian09:06
=== seb128 [~seb128@ANancy-151-1-11-54.w83-194.abo.wanadoo.fr] has joined #ubuntu-devel
Keybuk(though if you're lucky it's a silly whitespace change or something)09:07
=== g-remy [~chatzilla@lns-vlq-17f-81-56-171-105.adsl.proxad.net] has joined #ubuntu-devel
Kamionmdz: the string "udevd" doesn't appear anywhere in the initrd09:09
mdzKamion: er, so udevd isn't actually present?09:09
Kamion/sbin/udevd is there09:09
mdzthat counts as a string :-)09:09
Kamionno other occurrence of udevd in filenames though09:09
Kamioner, files09:09
mdzKamion: what's in /etc/hotplug.d?09:10
Kamionchandev default dock ieee1394 input net pci scsi usb09:10
=== Nigelenki [~bluefox@pcp485126pcs.whtmrs01.md.comcast.net] has joined #ubuntu-devel
Kamionah, default/10-udev.hotplug is a symlink to udevsend09:10
Kamionthat probably explains it09:11
mdzKamion: that's what I meant by: <mdz> Kamion: maybe the udev hotplug hook starts udevd if it isn't running?09:11
Kamionright, I missed it 'cos it was a symlink09:11
mdz       If udevd isnt already running, udevsend will start it.09:12
Kamionis there a way to tell it to queue events instead?09:12
mdzKamion: kill -STOP udevd? :-)09:12
Kamionno, I have to make it go away09:12
mdzKamion: you could disable hotplug09:12
Kamionit's stopping me umounting /initrd09:12
Kamionfrom where? :)09:13
Kamionrun from where?09:13
Kamionthis is *before /sbin/init*09:13
mdzseb128: here?09:13
mdzKamion: *blink*09:13
seb128mdz: yes09:13
Kamionmdz: 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 running09:14
mdzseb128: gnome-settings-daemon segfaults in hoary09:14
mdzseb128: is this expected?09:14
seb128mdz: no09:14
seb128let me upgrade my test box 09:14
Kamionmaybe I could install hotplug somewhere different and move it into place after pivot_root or something09:15
mdzseb128: /msg'd you the backtrace09:15
KamionI can't wait until anna09:15
seb128mdz: ok, thanks09:15
mdzKamion: is there a way to set a sysctl from the kernel command line?09:16
Kamionmdz: hm, the kernel command line is already running badly out of space for preseeders, don't want to stress it any more09:16
mdzKamion: you could just fuser -mk before unmounting it09:16
Kamionno fuser, and not sure I want to risk killing udevd in the middle of handling an event09:16
mdzif it's running in something that isn't the real root anymore, should be fairly harmless09:17
Kamionyeah, but it might mean some bit of hardware doesn't get recognised09:17
KamionI always thought boot events were queued09:17
=== justdave [~dave@] has joined #ubuntu-devel
mdzKamion: later on in the boot process, the udev init script tells it to create nodes for everything it already knows about09:19
KamionI think it's probably going to be more predictable to ensure hotplug is not run in the initrd09:20
mdzI imagine that in a normal system, the events are just ignored in the initrd, since it doesn't have hotplug09:20
=== nasdaq4088 [sdfsd@tkp-ip-nas-1-p232.telkom-ipnet.co.za] has left #ubuntu-devel []
seb128mdz: control-center crash fixed in 2.8.1-0ubuntu2 just uploaded09:26
Kamionmdz: lucky you backed out that perl-in-hotplug thing actually, otherwise trying to use it in d-i would be a total bust09:27
elmokamion: ah, yeah, damn, fair point09:28
KamionI suspect there'll be problems anyway, /etc/hotplug/*.rc uses a load of stuff that I bet isn't available in busybox09:29
Kamionelmo: since it needs to be available in both places, hardlinks/copies are probably the only option :(09:31
elmoyeah, but the whole point of pools was to avoid managing hardlink farms.  gar, sucks.09:31
Kamionyeah, I know :(09:31
Kamionmdz: 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
Kamionmodules.*map I guess09:33
Kamionarrrr, this sucks, they're huge09:33
mdzI didn't type that fast enough to evade workrave09:33
=== x4m [~max@230.237-200-80.adsl.skynet.be] has joined #ubuntu-devel
Kamionalthough only 63K compressed, I guess09:33
mdzI haven't figured out its algorithm yet, but it has a certain small tolerance built into it09:34
Kamionmaybe I can figure out a reduction algorithm; like, only include the lines that actually match modules available in d-i :-)09:35
Kamionthere is a problem with amd64: the boot messages are visible for about a tenth of a second :P09:37
mdzKamion: gzipping them wouldn't be a half bad idea, except that hotplug reads them about 80 BILLION TIMES09:37
Kamionmdz: don't care so much about uncompressed size for now, although the lowmem people will care09:38
mdzI'm surprised they don't compress better than that09:38
mdzmdz@potpal:/lib/modules/ $ cat *map |wc -c09:38
mdzmdz@potpal:/lib/modules/ $ cat *map |gzip -9 |wc -c09:38
Kamionoh, I was using modules.*09:39
mdznot entirely sure whether modules.alias is used still09:39
mdzmodules.dep you need anyway, of course09:39
KamionI think it09:40
Kamion's reduced, though09:40
mdzelmo: gnutls11 is built, but its binaries are in universe (though depended upon by packages in main)09:41
elmomdz: yeah09:41
seb128mdz: libxml++2.6 is new in deb and not in universe ... what should be done to get it, just ask a sync ?09:43
elmolibgcrypt11 would be another one to merge09:43
elmoseb128: we're missing new packages ATM, I'll hopefully fix that tonight09:44
seb128ok, thanks09:44
mdzelmo: libgcrypt11 is safe to overwrite with the version from Debian, please sync it09:45
elmocool, done09:45
elmothen I can move it to main ;-)09:46
=== sm [~simon@lsanca1.ar5-] has joined #ubuntu-devel
Keybuk\_ patch -stuN -z .magic-orig -r /home/scott/magic.patch-reject -p109:56
KeybukI love that command09:56
Keybuk"set patchers to stun"09:56
mdzKeybuk: you so added the -u just to make a word :-P09:58
=== Keybuk whistles and examines the ceiling
=== graham [~graham@capella.plus.com] has joined #ubuntu-devel
=== mdz_ [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #ubuntu-devel
grahamtseng, are you the tseng of mono repository fame?10:11
grahamwould you be interested another package? I've recently packaged Spam Trainer.10:11
tsengits not very useful to me10:12
tsengbut if it works, you can email it to me and ill upload it eventually10:12
grahamfair enough10:12
elmoWTF is with bugzilla and firefox?10:12
grahamincidentally, 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
elmoKamion: ?10:13
tsenggraham: there is a syntax error10:13
tsengmissing )10:13
grahamI 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
grahamdo you use pbuilder?10:13
tsengi use dpkg-buildpkg10:13
grahamokay. I'll check it out when I get my laptop out as I need to run spam trainer through pbuilder anyway.10:14
tsengyou can fix the package by fixing the ) in debian/control10:15
grahamok, thanks.10:15
=== amu is back from the LWE
amumark#s talk was nice, and ubuntu get into the final for an award  10:18
mdz_elmo: which bugzilla/firefox bit in particular?10:19
=== sabdfl [~mark@host217-37-231-28.in-addr.btopenworld.com] has joined #ubuntu-devel
amusabdfl: wb 10:19
mdz_elmo: you can file those bugs in bugzilla, some of them are already there and some have been fixed10:19
elmomdz: it keeps hanging when I scroll down with the wheely mouse button10:19
sabdflhey all10:19
mdz_elmo: ah, that one10:20
=== lamont is happy to announce that king is back in the rotation
sabdflhey amu, long time no see :-)10:20
mdz_elmo: happens with some other sites, too10:20
sabdfllamont: long live the...?10:20
lamontand we have buildd/sbuild for amd64 architectures now.10:20
mdz_elmo: https://bugzilla.ubuntu.com/show_bug.cgi?id=196210:20
amusabdfl: i joined the social event, ubuntu was was one of the final getting an award   10:20
lamonthoary.all.amd64:Total 858 package(s) in state Needs-Build.10:21
lamonthoary.all.i386:Total 1 package(s) in state Needs-Build.10:21
lamonthoary.all.powerpc:Total 123 package(s) in state Needs-Build.10:21
sabdflwho was issuing th awards?10:21
lamontamu: cool!10:21
sabdfl"best nekkid people"?10:21
amusabdfl: LWE for best distros's best database ..... 10:21
=== dante_ [~dante@exhax10-081.dialup.optusnet.com.au] has joined #ubuntu-devel
mdzwho was _receiving_ the award? :-)10:21
amuskole :) 10:21
Keybukelmo: where do you want a list of packages we can sync from Debian?10:21
Keybuk(because they took all our changes)10:21
mdzKeybuk: yes, he does :-)10:22
mdzamu: surely you're joking, Mr. Mueller10:22
elmokeybuk: mail or url link - whatever10:22
sabdflskole won an award, over ubuntu?10:22
amumdz: it's like this 10:22
Keybukelmo: http://people.ubuntu.com/~scott/revert10:23
Keybukthat's the first chunk10:23
amuguess they should add new new award, for the best debian distro10:23
mdzamu: confused10:24
mdzwho gave an award to which distribution, and who accepted it?10:24
amumdz: ... the last years all time debian won 10:24
elmoKeybuk: you're 100% sure about this?  i can't undo once I start10:24
elmo(sanely anyway)10:24
amuLWE gives some awards to the projects.  10:24
lamontelmo: and insanity takes is _soo_ painful.10:24
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
amuthe distro award goes to skole this year 10:25
Keybukelmo: yeah, they all resulted in a zero debian->hoary diff10:25
Keybuk(well, just changelog mogs)10:25
mdzamu: bu tyou said ubuntu received an award?10:25
mdzs/bu ty/but y/10:25
mdzKeybuk: did you eyeball them?10:26
Keybukmdz: yup10:26
Kamionelmo: yep?10:26
amumdz: one of the final canditates ... but skole won at least 10:26
mdzamu: ah, ok. so ubuntu was a finalist10:26
amumdz: yap 10:26
mdzbut did not receive an award10:26
amuyap 10:27
elmoKeybuk: ok, doing now10:27
grahamit appears as though "skjermbilder" means "screenshot" in norwegian. neat. I've never heard of skoke before; what did they win for?10:28
__danielgraham: good question :-)10:28
grahamI'm thinking it must be amazing...10:28
Kamionskolelinux are a Norwegian Debian derivative; they do mass rollouts of preconfigured systems with automatic LDAP setup etc. in schools.10:29
Kamionquite specialised10:29
Kamionany udev experts in the house?10:30
Kamionhow do I make it create /dev/fb/0?10:30
KamionI'm using devfs.rules and compat-full.rules10:31
Keybukwhat's it creating currently?10:31
__danielKamion: this is what i got in devfs.rules:  KERNEL="fb[0-9] *",  NAME="fb/%n",10:32
Kamion__daniel: yes, likewise10:32
KamionKeybuk: nothing for the framebuffer device that I can find10:32
KeybukKamion: do you have the appropriate module for it loaded?10:32
KamionKeybuk: yes10:32
Kamionvga16fb and fbcon10:32
Keybukand you're not getting either /dev/fb0 or /dev/fb/0 ?10:33
KamionI can make everything work if I create the device node by hand10:33
KamionKeybuk: nope10:33
elmoKeybuk: k, imported10:33
Kamionit's possible hotplug was hosed actually, I'll try again10:33
=== elmo goes to eat dinner - bbiab
Kamionminor detail of having the wrong thing in /proc/sys/kernel/hotplug ...10:34
Keybukelmo: imported what where?10:34
=== Kamion suspects he means your key
__danielKamion: works at my place... ah ok10:35
Kamion__daniel: this is in the installer FWIW, it's a very minimal system and entirely possible I've left bits out :)10:35
Kamiontrying to figure out the chain of code involved10:35
__danielKamion: oh... i see10:36
Keybuk98819394 bytes transferred in 2 seconds (62.16M/s)10:36
KeybukTotal 203 files transferred10:36
Keybuklet's see if I get any messages10:39
=== lamont wanders off for a few minutes
KeybukKamion: I don't think he did import my key10:49
Keybukno response yet10:49
elmono, not that10:50
elmothe all-merged-up stuff is imported from sid10:50
Kamionsorry, misinterpreted10:50
Kamiongrr. There's a delay between a module being loaded and the device node being visible, isn't there?10:51
Keybukelmo: see, now you have 203 files in the queue to get rid of as well as import my key <g>10:51
KeybukKamion: yeah, but usually not a huge one10:51
KamionKeybuk: is it predictable in any way?10:52
Keybukno more than a second or two10:52
KamionI don't want to insert random sleeps, that sucks10:52
KeybukKamion: no, it's basically just how long it takes for the stuff in /sys to appear10:52
Kamionis there any way to say "wait until everything's synced"?10:52
Keybukhotplug is full of "wait for file to appear" checks10:52
Kamionthat's really crap10:52
Keybukfor which?10:52
Keybukif you're doing a hotplug event, things will be synced at that point10:53
Kamionmodprobe fbcon -> wait for /dev/fb/0 to show up10:53
Kamionin a sane world you can expect it to be there from the point modprobe exits10:53
Kamionit'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 appear10:54
Keybukthat'd mean it'd all have to happen in kernel space10:54
KamionI don't care how it's implemented :)10:54
KamionI just want a supposed upgrade not to be a regression10:54
Keybukwouldn't doing the modprobe populate something under /sys ?10:54
Keybukno :p10:55
Kamionalso, sleeps are a totally obvious race condition10:55
Kamiond-i runs on some very slow hardware10:55
Keybukthe trouble is you're doing it backwards, really10:56
Keybukyou're not supposed to load a module to find out if the hardware exists10:57
Keybukyou're supposed to load the module once you know the hardware exists10:57
Kamionis there a way to test for available framebuffer hardware, then?10:57
Keybuk(this obviously isn't quite finished yet, as things like ppp_generic/loop go phooey)10:58
Keybukwell, the graphics card's pci id will map to the appropriate frame buffer10:58
Kamionall I want, really, is a wait_for_everything_to_sync call10:58
Keybukradeonfb             0x00001002 0x00005961 ...10:59
Keybukfor example10:59
Kamionsounds a bit challenging to parse at this stage, but I guess10:59
Kamionah well, that will come later, sleep it is for now11:01
Kamionthanks for guidance :)11:02
Keybuknite :)11:02
__danielsleep tight, kamion11:02
Kamionsleep> i.e. /bin/sleep not bed11:02
Keybukoh, lol11:02
Keybukmy brain saw "sleep" and thought "mmm... bed"11:03
__danielme too11:03
Keybukdunno why, but I've been up at 5am every morning this week11:04
=== amu mv /dev/bed n8
=== Keybuk thinks he'll do the same in a minute
=== azeem [~mbanck@socks-out.lrz-muenchen.de] has joined #ubuntu-devel
Kamionudev/hotplug d-i works11:42
Kamionthough I haven't tried to strip out discover yet11:42
jdubarh, no keybuk11:47
jdubwhat's the best way for me to upload a modified/updated gdm?11:48
jdubjust grab the new release, apply changes and go?11:48
seb128good luck11:49
seb128I've tried once11:49
seb128neuro has all his changes mixed in the diff.gz without any documentation11:50
=== jdub will attempt :)
seb128Np237 took a whole week to repackage it with cdbs last year11:50
jdubi'm tempted to fork it ;)11:50
seb128use Np237's work if you do this11:50
seb128he made a good package using cdbs with all the patches splitted in debian/patches11:50
seb128jdub: hum, we want glib/gtk 2.5 in hoary now, right ?11:52
seb128ok, that's a bit late now to start packaging them, but that's on my todo list for tomorrow :)11:53
seb128jdub: if you want the gdm package made by Josselin you can find it in the pkg-gnome SVN (/packages/gdm)11:55

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!