[12:08] <bddebian> Howdy
[12:19] <jdong> is gnome-panel cpu hogging for anyone else?
[12:19] <jdong> it chews a constant 70%-ish for me
[12:19] <jdong> (edgy)
[12:20] <shining> see bugs
[12:21] <shining> I don't know what's the status how it though, I've read some stuff on ubuntu, debian and gnome bugzilla
[12:24] <jdong> btw, when'd /bin/sh turn into dash?
[12:24] <theCore> jdong: it did on my system once, I had to `killall gnome-panel' to fix it
[12:25] <jdong> theCore: it happens every login every time on my system right now :-/
[12:25] <jdong> doing a dist-upgrade right now, but don't see anything that should affect the panel
[12:25] <crimsun> is it reproducible for a brand new user?
[12:26] <shining> erm, I'm looking at the bugs again, some of them are a few month old, and are marked as fixed
[12:27] <jdong> crimsun: happens in my freshly upgraded edgy
[12:27] <jdong> when I get a chance, I'll try a livecd
[12:27] <shining> I just removed a broken symlink
[12:27] <jdong> but it's got me running kubuntu for now ;)
[12:27] <shining> or kill gamin
[12:28] <jdong> hmm, kill gamin.... that sounds like an idea
[12:28] <shining> hehe, me too :)
[12:30] <shining> it's the second bug on gnome-panel
[12:30] <shining> https://launchpad.net/distros/ubuntu/+source/gnome-panel/+bugs
[12:31] <shining> high and confirmed
[12:31] <shining> there are at least three workarounds
[12:31] <shining> killing gamin, removing the broken link, fixing the link (by installing menu or something)
[12:32] <jdong> shining: thanks, will do workarounds, but that still doesn't address the issue :-/
[12:32] <shining> no
[12:32] <shining> since it's a problem with gamin, either that needs to be fixed or something else used instead
[12:33] <shining> also, fixing the menu packages to not provide a broken symlink wouldn't hurt, I think
[12:36] <jdong> BenC: just want to tell you edgy kernel is a charm on my core duo
[12:36] <jdong> kudos
[12:37] <shining> maybe also check these:
[12:37] <shining> https://launchpad.net/distros/ubuntu/+source/gamin/+bug/36581
[12:37] <Ubugtu> Malone bug 36581 in gamin "gam_server consumes lots of cpu time" [High,Needs info]  
[12:37] <shining> http://bugzilla.gnome.org/show_bug.cgi?id=323064
[12:37] <Ubugtu> Gnome bug 323064 in general "the panel applications menu is displayed empty with gamin 0.1.7/inotify" [Normal,Unconfirmed]  
[12:42] <jdong> whoa, is edgy gcc stack protector'ed?
[12:43] <tseng> yes.
[12:43] <jdong> cool
[12:44] <jdong> is there any work towards other forms of security?
[12:44] <jdong> like exec-shield/selinux/whatever-else-redhat-is-advertising?
[12:44] <tseng> no.
[12:45] <jdong> k. I like the brief answers
[12:45] <jdong> will I get rich today?
[12:45] <tseng> nope.
[12:45] <jdong> will you ever be more optimistic?
[12:45] <bddebian> heh
[12:45] <tseng> no.
[12:45] <jdong> where's my firefox 2.0b2 packages?
[12:46] <tseng> iwj is on vacation.
[12:46] <jdong> magic 8 ball, will applying 2.0b1 diff.gz to 2.0b2 sources work?
[12:46] <tseng> nope.
[12:46] <jdong> tseng: is that for real, or are you just trying to be funny now?
[12:46] <tseng> ok, "extremely unlikely"
[12:47] <jdong> lol, I don't doubt it
[12:47] <tseng> it usually takes days to work it out
[12:47] <shining> really?
[12:47] <shining> I wouldn't have thought so
[12:47] <shining> what does generally raise problems?
[12:47] <jdong> I'll have an answer in 45 seconds
[12:47] <jdong> we do have a bunch of patches
[12:48] <jdong> which will likely reject like hell
[12:48] <tseng> shining: a large set of patches against a large code base
[12:48] <tseng> with a large delta of its own
[12:48] <shining> ah ok
[12:48] <jdong> 10 seconds...
[12:48] <jdong> wow, I love these perpendicular recording hard drives
[12:48] <jdong> alright, all unpacked
[12:48] <jdong> now, for the moment of truth......
[12:48] <shining> I'm not sure if I need all these patch though
[12:49] <bddebian> Hmm, is the cvs stuff on cvs.fedora.redhat.com just their patches?
[12:49] <jdong> failed hunks
[12:49] <jdong> tseng is right :)
[12:49] <jdong> LOL, my friend just im'ed me asking what all that disk grinding was :P
[12:50] <shining> oh, so that was for real :d
[12:50] <jdong> what, you think I was kidding?
[12:50] <rshakin> hello guys 
[12:50] <shining> yep 
[12:50] <jdong> I built that conroe extreme in return for free ssh access :)
[12:50] <jdong> I can't believe the guy actually agreed :P
[12:50] <shining> :d
[12:51] <rshakin> sorry to bother you ppl but i need some help 
[12:51] <rshakin> seems to me that the folks in #ubuntu are ether ignorinng me or something...
[12:51] <jdong> rshakin: shoot. it better be good or we'll ignore you too ;)
[12:52] <jdong> if you don't ask quickly, someone will hurry you off to #ubuntu again :)
[12:53] <shining> well, it has been off topic for a few hours, so I guess it doesnt matter
[12:53] <rshakin> ok, i am getting a shitload of erors when i try to update my system...  http://madwifi.org/paste/161
[12:53] <rshakin> thats what i get when i try to update using apt-get update 
[12:53] <jdong> you spelled archive wrong :)
[12:54] <jdong> that's your first problem
[12:54] <shining> lol, so lame
[12:54] <jdong> and it looks like you've got some connection problems to us.archive.ubuntu.com, too
[12:54] <jdong> since I'm using that mirror right now to download 400MB of updates, I'd say your network connection is wacky
[12:55] <rshakin> hmm weird 
[12:55] <rshakin> i am on it right now :) 
[12:55] <jdong> either way, fix that typo first :)
[12:55] <rshakin> hehe... will do that 
[12:55] <bddebian> doko: You aren't around by any chance are you?
[12:56] <jdong> bddebian: wow, still working hard on azureus?
[12:56] <rshakin> btw this is a clean install from a cd... nothing on it except what came with 
[12:58] <rshakin> this is just weird it's like my isp is  blocking that site 
[12:58] <bddebian> jdong: Just now getting to it
[12:58] <jdong> rshakin: well, I don't believe that archive was spelled wrong on the cd.... but your isp might have connection issues to the mirror
[12:59] <jdong> bddebian: k, cool... fedora's azureus runs really well with gcj, so let's hope we're that lucky too
[12:59] <rshakin> well the archive was my fault... i was trying to edit to another mirror... so yeah thats more of my fault 
[01:00] <jdong> :)
[01:00] <rshakin> but as far as mirror, you are right there is no traffic going through to it for some off reason... 
[01:00] <bddebian> jdong: Well I don't see the actual source.  Pulling from redhats repo all I seem to get are the patches
[01:00] <jdong> bddebian: look for fedora extras' src.rpm's then?
[01:00] <rshakin> is there another mirror that i can use  besides us.archive 
[01:00] <pygi> rshakin, yes
[01:01] <bddebian> rshakin: Just make it archive.ubuntu.com
[01:01] <jdong> tseng: oh mono god, is there any way to get gtk# docs in monodevelop's help panel?
[01:02] <rshakin> hmm, this is just plain weird root@rshakin-laptop:/home/rshakin# apt-get update
[01:02] <rshakin> Err http://archive.ubuntu.com dapper Release.gpg
[01:02] <rshakin>   Connection failed [IP: 195.248.90.23 80] 
[01:02] <jdong> rshakin: are you behind a proxy or anything like that?
[01:02] <jdong> rshakin: did you install a firewall, etc?
[01:02] <tseng> jdong: right, you got it
[01:03] <rshakin> jdong: router wrt51G going to reset it in a min to see if it helps 
[01:03] <jdong> tseng: wow, mono really rocks
[01:03] <pygi> jdong, :)
[01:03] <tseng> yeah, nice language
[01:03] <hikenboot> hello all--I asked in #ubuntu but will ask here ...is there a list somewhere of packages that are installed in ubuntu live cd 6.06.1 that are not necessary for the basic function of the cd ..I am trying to make a custom live cd and want to make room for my own packages by removing all workstation related packages except gnome itself
[01:04] <tseng> gtk# for the win
[01:04] <pygi> hikenboot, you have several "modify livecd" apps
[01:04] <jdong> tseng: do you know how well mono would work on arm?
[01:04] <tseng> jdong: it works on arm...
[01:04] <tseng> arm-l support is new
[01:04] <tseng> i cant get it going on 770 on first try
[01:04] <tseng> someone apperantly has, havent had much time for it
[01:04] <jdong> tseng: there's the JIT compiler on arm, right?
[01:05] <jdong> does mono's jit compiler generally make the code execute faster than python?
[01:05] <tseng> generally faster than python...
[01:05] <bddebian> Egads, how do I extract a clean gcj to make a patch?
[01:05] <tseng> thats pretty hard to answer
[01:06] <jdong> k
[01:06] <Burgwork> hikenboot, check the seeds
[01:06] <jdong> tseng: grr, monodevelop doesn't let me cut and paste from the documentation browser? :)
[01:06] <hikenboot> pygi, besides livedistro what others got a link to a list somewhere?
[01:06] <tseng> jdong: why wouldnt it?
[01:06] <hikenboot> check the seeds? what do you mean?
[01:06] <tseng> jdong: i just did
[01:07] <jdong> tseng: almost everything under edit is greyed out while the documentation tab is active?
[01:07] <pygi> hikenboot, I'll give you a script that you'll run
[01:07] <tseng> i went back to the file tab
[01:07] <pygi> hikenboot, sec pls
[01:07] <tseng> and pasted
[01:07] <Burgwork> hikenboot, ubuntu is built from seeds. see people.ubuntu.com/~cjwatson/seeds/
[01:07] <pygi> hikenboot, COLUMNS=200 dpkg -l | awk ' { if ($1 == "ii") print $2; }' > /tmp/pkglist.txt
[01:08] <jdong> tseng: middle-click pasting works, but the menu options are all disabled... can you confirm that?
[01:08] <pygi> hikenboot, my command should also help :P
[01:08] <tseng> i dont use anything but "middle click pasting"
[01:08] <tseng> what menu do you mean, exactly
[01:08] <jdong> tseng: edit->copy
[01:08] <jdong> it's greyed out in the documentation tab
[01:09] <tseng> ok, yes
[01:09] <jdong> :-/ I'd call that a missing feature :)
[01:10] <hikenboot> hey pygi, thats great...its a much shorter list than i generated with other commands i have used..I assume these are the root packages that install the others?
[01:10] <tseng> I personalyl would have never noticed
[01:10] <tseng> the documentation tab is embedding monodoc
[01:10] <jdong> hmm
[01:11] <jdong> does stuff written on edgy in mono typically work in dapper?
[01:11] <jdong> i.e. how's mono's binary compatiblity?
[01:11] <pygi> hikenboot, those are all packages you have
[01:11] <tseng> it should be good
[01:11] <tseng> back a few versions
[01:11] <tseng> except for say, sharpzlib
[01:11] <tseng> that was a nice abi break 
[01:11] <hikenboot> ah instead of package i guess i was showing deb's or somthing
[01:12] <hikenboot> thanks a bunch guys
[01:12] <jdong> tseng: how's running these exe's under windows?
[01:12] <tseng> jdong: not good
[01:12] <tseng> libraries
[01:12] <jdong> ah, need gtk#?
[01:12] <tseng> you can get gtk#
[01:12] <tseng> if you use gnome stuff, it gets tough
[01:13] <jdong> does gtk# run under ms's .net?
[01:13] <tseng> yes
[01:13] <bddebian> Scary :-)
[01:13] <jdong> k, so just need to get people to find gtk# :-/
[01:13] <jdong> how's windows.forms under linux?
[01:13] <tseng> pretty good
[01:14] <jdong> is there any gui SWF designers for linux?
[01:14] <tseng> i dont think so
[01:14] <jdong> well, that ruins that idea :-/
[01:15] <jdong> well, that concludes my mono interrogation session for today
[01:15] <jdong> thanks for your time, tseng
[01:16] <tseng> that was easy
[02:36] <robertj> is Trash a place in edgy now?
[02:38] <tseng> no
[04:58] <msw> jdub: dude, you're jumping to conclusions
[05:17] <nixternal> OOo broke?
[05:19] <bluefoxicy> OOo shiny
[05:20] <nixternal> heh
[05:20] <nixternal> OOo dead on one machine, slow on the other
[05:21] <nixternal> splash screen pops up, and then disappears right away =/
[06:04] <RedRose> toshiba: not a supported Toshiba laptop It recommend I rebuilt My kernel, should I, and how?
[09:12] <Mempf> a lot of my fonts in firefox or err bon echo are ugly
[10:46] <hunger> What is the recommendation on upstart use? I read it should be tested side by side with sysvinit, but now those two conflict.
[10:46] <giftnudel> hunger: just uninstall sysvinit ;)
[10:47] <hunger> giftnudel: That will breaks ubuntu-minimal.
[10:47] <giftnudel> hunger: really?
[10:47] <giftnudel> hunger: for me it didnt
[10:47] <_ion> It doesn't "break" it, it just removes it. :-)
[10:48] <hunger> _ion: Well, OK, I should have explained that better;-)
[10:48] <_ion> No harm done. Install ubuntu-minimal later again, when it allows upstart.
[10:48] <giftnudel> hunger: are you sure it breaks it? It only uninstalled ubuntu-minimal for me
[10:48] <hunger> giftnudel: s/breaks/uninstalls/
[10:48] <giftnudel> ehm, it only uninstalled sysvinit for me
[10:48] <giftnudel> hunger: forget the statement before the last
[10:49] <giftnudel> that was wrong
[10:49] <hunger> Sorry, I should have said the right thing instead of claiming breakage.
[10:49] <giftnudel> hunger: so correctly: I did not need to uninstall ubuntu-minmal
[10:50] <hunger> Hmmm... aptitude is not happy with my current setup anyway, wanting to uninstall ubuntu-desktop due to broken dependencies all the time.
[10:50] <giftnudel> I just did apt-get install upstart and it only removed sysvinit
[10:50] <hunger> Uninstalling ubuntu-minimal should not make matters worse:-)
[10:50] <giftnudel> hehe
[10:51] <zyga> hunger: upstart works okay
[10:51] <zyga> just make sure you got upstart-sysvinit-compat 
[10:51] <hunger> zyga: Yeap, so it does for me.
[10:51] <zyga> without that you're hosed :)
[11:04] <zyga> hmm
[11:04] <zyga> can data.tar.bz2 be used in debian packages?
[11:06] <zyga> anyone?
[11:13] <hunger> It would have been nice if somebody had warned me that updating upstart causes kernel panics:-|
[11:13] <_ion> hunger: It shouldn't  or maybe it does if you upgrade from < 0.2...
[11:14] <hunger> _ion: I assumed that it should not;-) What should cause kernel panics after all?
[11:14] <_ion> hunger: init (pid 1) dying.
[11:15] <hunger> I still find it pretty scary that upstart is supposed to replace inetd.
[11:15] <hunger> _ion: You find it great that process 1 listens on the network?
[11:15] <_ion> It doesn't have to be process 1.
[11:16] <hunger> Oh, then it is closer to being OK...
[11:16] <_ion> For all i know, it could be a helper process that communicates with init.
[11:26] <Hobbsee> Kamion: drat.  so i cant use requestsync script on it's own then :(
[11:33] <giftnudel> _ion: well, where did you get upstart-sysvinit-compat from?
[11:33] <hunger> giftnudel: apt-get install?
[11:33] <giftnudel> in edgy?
[11:33] <hunger> giftnudel: worked for me(TM)
[11:33] <hunger> giftnudel: Yeap.
[11:33] <giftnudel> what?
[11:33] <_ion> Yes.
[11:34] <giftnudel> maybe I should check my sources.list
[11:34] <ajmitch> upstart-compat-sysv
[11:34] <giftnudel> but I don't have it ...
[11:34] <ajmitch> not upstart-sysvinit-compat
[11:34] <giftnudel> ah, that may have been the cause ;)
[11:34] <giftnudel> now i find it ;)
[11:35] <giftnudel> thanks ajmitch
[11:53] <lasindi> Hi everyone, I'm trying to create my own custom Ubuntu live CD, and I'm following this tutorial: http://www.atworkonline.it/~bibe/ubuntu/custom-livecd.htm I've written a script that I'd like to create a shortcut on the desktop to (like the "Install" shortcut to the installer). Since this shortcut must be copied to /home/ubuntu/Desktop at boot, is there a script in the CD that creates the ubuntu user that I can modify?
[11:55] <zyga> _ion: install upstart-sysv-compat
[11:56] <zyga> or similar..
[11:56] <zyga> and make sure to REMOVE the alternative upstart section from menu.lst
[12:05] <pygi> zyga, upstart-sysv-compat package isn't there
[12:05] <pygi> ah :P
[12:05] <hunger> pygi: Try upstart-compat-sysv (apt-cache search upstart will list it).
[12:05] <pygi> hunger, I am running upstart already, no worries :)
[12:06] <hunger> pygi: Good. Did you get a kernel panic upgrading it?
[12:06] <pygi> hunger, nop, all worked
[12:06] <hunger> pygi: Just asking, because I got one:-(
[12:26] <_ion> Haha, pipermail ftw. http://lists.netsplit.com/pipermail/upstart-devel/2006-September/000006.html
[12:53] <pygi> hey ho siretart 
[01:27] <zyga> sigh
[01:28] <zyga> python's tarfile module is broken
[02:20] <_ion> Could fi.archive.ubuntu.com be changed to point to 193.166.3.2 (the address of ftp.funet.fi)? It's a very fast mirror for Finnish users.
[02:23] <Lathiat> _ion: i beleive one of the funet admins would need to approac the ubuntu-mirrors team?
[02:23] <Lathiat> i could be wrong someone else may know better
[02:25] <Lathiat> well i guess the ubuntu guys could be pursuaded ot harass the funet admins :)
[02:26] <_ion> (The mirror is already there in the correct path, http://ftp.funet.fi/ubuntu/)
[02:28] <_ion> elmo: Any comments?
[02:32] <Seveas> _ion, #ubuntu-mirrors
[02:32] <_ion> seveas: Thanks.
[03:06] <lisi> mjg59, hello, ping, here ?
[04:01] <sladen> 650MB for Firefox(!)
[04:01] <pygi> sladen, tabs? :P
[04:10] <sladen> pygi: "only a few" :)
[04:13] <sladen> ...dozen.
[04:14] <zyga> re
[04:14] <pygi> sladen, :P
[04:20] <JanC> weird, synaptic (in dapper) acts differently (and very strange) when I manually mark the new OOo packages from dapper-proposed for upgrade (a.o. it wants to install kdelibs on my GNOME system???), compared to what happens when I mark them with "mark all upgradeable packages for upgrade" (does what I expect it to do)
[04:20] <JanC> is that a bug or somehow a feature  ;-)
[04:20] <Hobbsee> JanC: to install kde stuff on your machine?  of course it's a feature!
[04:21] <Hobbsee> JanC: one of the early steps for the plan of world domination, by KDE.
[04:21] <Hobbsee> JanC: next step will be installing kubuntu-desktop at the same time.
[04:21] <zyga> lol
[04:21] <JanC> fortunately I always check what gets installed beforehand  :-P
[04:23] <zyga> I wonder what triggered the existence of mac forge
[04:23] <Lathiat> everyone needs their own forge
[04:24] <zyga> it's pretty useless ATM: they push bonjour, we have avahi, they push launchd, wh've just got upstart
[04:24] <Lathiat> probably an apple-promotin-community-dev thing i would suspect
[04:24] <Lathiat> zyga: sure, but theyre still open to do that
[04:24] <zyga> Lathiat: it doesn't look like community, just official apple stuff
[04:24] <sladen> JanC: if its replicatable, time to file a bug
[04:24] <sladen> JanC: what happens from 'apt-get' or 'aptitude'
[04:25] <LarstiQ> zyga: to be fair, launchd exists quite a while
[04:25] <zyga> LarstiQ: yes I know
[04:25] <zyga> and it had a license change along the way
[04:25] <Lathiat> bonjour is still usefull on windows too
[04:25] <zyga> what I'm trying to say: it's not about creating a community with their own projects - you canont do that on mac forge
[04:25] <Lathiat> avahi doesnt cover that yet unfortunately
[04:25] <_ion> When the launchd license change happened, upstart was already better more suitable for Ubuntu.
[04:26] <JanC> apt-get does the right thing, as well as synaptics "mark all upgradeable" button (both in normal & smart mode) 
[04:26] <Lathiat> im pondering goign to the dark side to port avahi to windows
[04:26] <Lathiat> could be an interesting experience
[04:26] <_ion> s/better //
[04:27] <Lathiat> bbl
[04:27] <zyga> _ion: yeap, upstart is far from being finished but it's already more like what we need
[04:51] <giftnudel> anyone here how can rebuild python-mysqldb on edgy, compare bug 55047, it has been open for nearly a month now (or are there any other problems with this package?)
[04:51] <Ubugtu> Malone bug 55047 in python-mysqldb "python-mysqldb doesn't actually install any python" [High,Confirmed]  http://launchpad.net/bugs/55047
[05:07] <jdong> how do I get the uuid of a disk?
[05:07] <jdong> is that something that changes if I mkfs?
[05:09] <_ion> jdong: vol_id(8) or ls -l /dev/disk/by-uuid, and yes.
[05:10] <_ion> Actually, the UUIDs are for partitions, not HDDs.
[05:10] <darius_> Can I upgrade to Edgy or does it have to be a fresh install?
[05:11] <_ion> Naturally you can.
[05:11] <jdong> _ion: so, if I just mkfs'ed my /dev/sda1 and rsynced a backup back onto it, do I have to change anything?
[05:11] <jdong> would the uuid have changed and I need to check my fstab/menu.lst?
[05:11] <_ion> jdong: Better to check fstab and menu.lst.
[05:12] <darius_> The 5.04 to 6.06 upgrade path presented me with a single dialogue to upgrade my distribution.  When I changed /etc/apt/sources.list to edgy it gives me 174 packages.  Not sure what I should expect
[05:12] <jdong> _ion: yep, it did change
[05:16] <rod> do the latest nvidia drivers support xorg7.1?
[05:17] <shining> they say so
[05:19] <geser> giftnudel: I've a rebuild version of python-mysqldb. could you test it?
[05:19] <rod> so when using nvidia, a simple apt-get install compiz will be sufficient to test aixgl and compiz?
[05:19] <shining> preparing to crash my laptop by trying to hibernate
[05:19] <shining> rod: there is no such thing as aixgl
[05:20] <rod> you got me confused
[05:20] <giftnudel> geser: sure
[05:20] <geser> do you have an amd64 or i386?
[05:21] <giftnudel> i386
[05:21] <giftnudel> geser: i rebuild one according to the bug (I added a comment)
[05:21] <giftnudel> this one worked
[05:21] <geser> I took the latest debian revision which also closes an other bug
[05:22] <giftnudel> geser: do you need my email or will you give me a link?
[05:23] <geser> I rebuild it now for i386 and then attach both debs to the bug report
[05:24] <giftnudel> ok
[05:24] <shining> crash succesful
[05:30] <tseng> Mithrandir: openbox 3.3. yay!
[05:32] <jdong_> did networkmanager break?
[05:32] <jdong_> ** (nm-applet:5672): WARNING **: <WARNING>       nma_dbus_init (): nma_dbus_init() could not acquire its service.  dbus_bus_acquire_service() says: 'Connection ":1.9" is not allowed to own the service "org.freedesktop.NetworkManagerInfo" due to security policies in the configuration file'
[05:50] <geser> giftnudel: I've attach deb for i386 and amd64 to bug 55047
[05:50] <Ubugtu> Malone bug 55047 in python-mysqldb "python-mysqldb doesn't actually install any python" [High,Confirmed]  http://launchpad.net/bugs/55047
[05:55] <giftnudel> geser: well, my program still works, but it doesn't use much of the module ;)
[05:59] <geser> thanks for testing
[06:01] <trappist> if a very minor bug is known to exist upstream, is it best to submit the fix upstream and wait for it to come back down the pike?  if so, what to do with the launchpad bug?
[06:02] <tseng> you can link the bug to upstreams bug tracker if its a common one
[06:02] <bddebian> Morning
[06:02] <tseng> if you cant you can mention the ustpream bug in a comment
[06:03] <tseng> don't crosspost, please
[06:03] <trappist> tseng: yes, sorry.
[06:08] <pixelmonkey> I just tried downloading the source for gnome-power-manager using apt-get source and then I tried building it using dpkg-buildpackage, but it claims unment dependencies.  When I try to apt-get the dependencies, it can't satisfy them... for example, libgnomeui-dev needs libgnomeui-0 of a version that won't be installed.  Shouldn't a clean dapper install, like mine, be able to build source packages?
[06:10] <Mithrandir> tseng: yeah, but it requires somebody to ask for a sync ; would you want to?
[06:10] <giftnudel> pixelmonkey: with apt-get build-dep gnome-power-manager, did you do that?
[06:10] <pixelmonkey> yea, it claims it can't satisfy them
[06:10] <tseng> Mithrandir: sure.
[06:11] <giftnudel> pixelmonkey: I think it's ok to open a bug o gnome-power-manager about that issue
[06:12] <_ion> Hi Keybuk
[06:12] <trappist> or on libgnomeui-dev if it's what has the unsatisfiable dep
[06:12] <_ion> keybuk: I'll write some tests for the patch.
[06:13] <pixelmonkey> giftnudel, see, the problem is that the dependencies are a little strange.  libgnomeui-dev expects 2.14.1-0ubuntu2, and indeed that is my version, but it also provides ubuntu3, which is what seems to mess up the install
[06:13] <Keybuk> _ion: cool :)
[06:13] <pixelmonkey> because apt-get says that 2.14.1-0ubuntu3 is to be installed, yet libgnomeui-0 provides both those versions
[06:14] <Keybuk> _ion: the test style should be pretty easy to copy
[06:15] <giftnudel> pixelmonkey: uh, well ... I don't know anything about that
[06:15] <Chipzz> pixelmonkey: are you sure you have a sane sources.list?
[06:15] <Mithrandir> pixelmonkey: you're aware that versioned provides are not implemented?
[06:15] <Chipzz> pixelmonkey: like, only entries for dapper-* and not for edgy?
[06:16] <zyga> Keybuk: did you notice upstart-initialized boot sequence to kill X soon after login and restart gdm normally afterwards?
[06:16] <Chipzz> pixelmonkey: also, apt-cache policy libgnomeui-0 might explain some stuff
[06:17] <giftnudel> Keybuk: do you know about problems with upstart and e2fsk after a crash?
[06:19] <pixelmonkey> Chipzz: http://paste.ubuntu-nl.org/22338 <-- not exactly sure what that means
[06:19] <Keybuk> zyga: no, I've not noticed this
[06:20] <Keybuk> giftnudel: no?
[06:20] <Chipzz> pixelmonkey: what it means is:
[06:20] <Chipzz>      2.14.1-0ubuntu2 0
[06:20] <Chipzz>         500 http://archive.ubuntu.com dapper/main Packages
[06:20] <tseng> Mithrandir: sync request filed
[06:20] <Chipzz> libgnomeui-0, if installed by apt, would be version 2.14.1-0ubuntu2, and it would be installed from that archive
[06:20] <Chipzz>  *** 2.14.1-0ubuntu3 0
[06:20] <Chipzz>         100 /var/lib/dpkg/status
[06:20] <Chipzz> this means
[06:21] <zyga> Keybuk: it has happened regularly before upstart completly replaced init, I did't notice any problems since
[06:21] <Chipzz> libgnomeui-0 is version 2.14.1-0ubuntu3, and it is only mentioned in that file (which is dpkg' status file)
[06:21] <giftnudel> Keybuk: I had xorg crash on me and after a reboot, I didn't have X there and logged into the console, I found out that e2fsk was running and apparently corrected some errors, and the system rebooted without a warning
[06:21] <Chipzz> it also means that the origin of the package is unknown
[06:22] <Keybuk> giftnudel: that's kinda kooky
[06:22] <pixelmonkey> Chipzz, hmm, I knew I shouldn't have run EasyUbuntu... I bet they grabbed it for some reason
[06:22] <Chipzz> pixelmonkey: ergo, you installed it manually from somewhere, or it got installed from an entry in sources.list which you have since erased
[06:22] <Keybuk> zyga: umm, so you're saying upstart fixed the problem?!
[06:22] <Chipzz> pixelmonkey: ah there you are
[06:22] <Chipzz> pixelmonkey: so, no bug
[06:22] <pixelmonkey> Chippz, right, I didn't think it was a bug
[06:22] <Chipzz> it's as simple as that
[06:23] <Chipzz> pixelmonkey: there is also nothing wrong with gnome-power-manager build-deps
[06:23] <pixelmonkey> Chipzz, what's a good approach I can take for reverting to the official package... I haven't dealt with apt hell fo awhile
[06:23] <Chipzz> since you are the one having unsupported packages installed
[06:23] <Keybuk> giftnudel: the console appears very early at the moment -- far earlier than it should, in fact
[06:23] <giftnudel> Keybuk: well at least at the next login, everything was working normally
[06:23] <Keybuk> it may be that you logged in while rcS was still running, which was why you didn't see X
[06:23] <Keybuk> don't know why it would reboot though
[06:24] <giftnudel> Keybuk: well, I had 3 getty's running (or that is as much as I tested)
[06:24] <Chipzz> pixelmonkey: I think apt-get install libgnomeui-0=2.14.1-0ubuntu2
[06:24] <Chipzz> or something like that
[06:24] <pixelmonkey> Chippz, cool thanks
[06:24] <giftnudel> Keybuk: the think is, it wanted to correct errors on the filesystem but apparently didn't wait for e2fsk
[06:24] <giftnudel> *thing
[06:24] <Chipzz> pixelmonkey: you may also have to downgrade other packages
[06:25] <Chipzz> pixelmonkey: but
[06:25] <Chipzz> pixelmonkey: do the following in a terminal
[06:25] <Keybuk> giftnudel: it could be that the checkfs script didn't manage to run sulogin properly?
[06:25] <Keybuk> but I thought then it'd just carry on with the boot sequence
[06:26] <Chipzz> pixelmonkey: zless /usr/share/doc/libgnomeui-0/changelog.Debian.gz
[06:26] <giftnudel> Keybuk: well, i was at terminal 7 and saw something similar to "something killed at vt2, restarting", then I switched to other consoles and logged in to see what wsa wrong
[06:26] <Chipzz> the entry at the top should tell you the changes and who uploaded it, and to where
[06:27] <pixelmonkey> "dapper-updates", by Sebastien Bacher
[06:27] <Chipzz> pixelmonkey: and do you still have dapper-updates in sources.list?
[06:27] <Chipzz> or did you remove that?
[06:27] <pixelmonkey> Chipzz, don't think so, never added it personally
[06:27] <pixelmonkey> Chipzz, though I have a feeling EasyUbuntu did :)
[06:28] <Chipzz> may very well be the case
[06:29] <zyga> Keybuk: no, I'm saying that upstart+compat-stuff finally stopped
[06:29] <zyga> it started right after installing upstart initially
[06:29] <pixelmonkey> Chipzz, ahhh.. I see... dapper-updates for some reason had its deb-src entry in my sources.list, but not the actual entry
[06:30] <pixelmonkey> Chipzz, looking at my sources.list file, it looks like it's been "script modified" a few too many times.  Last time I take that risk.
[06:30] <giftnudel> pixelmonkey: that explains the problem
[06:30] <pixelmonkey> I've been out of the loop awhile -- is dapper-updates a bleeding edge repository, or?
[06:31] <Chipzz> no no not at all
[06:31] <pixelmonkey> so it should actually be in my sources.list then?
[06:31] <giftnudel> not really, dapper-backports is more bleeding edgy, updates is for bugfixes and so on
[06:31] <pixelmonkey> ah, I see.
[06:31] <giftnudel> "bleeding edgy" ... ;)
[06:32] <pixelmonkey> I stop using computers for 8 months, get home and everything's different.  What a surprise :-/
[06:32] <Chipzz> not at all?
[06:34] <sladen> pixelmonkey: been locked up?
[06:34] <pixelmonkey> sladen, no no, I was traveling
[06:35] <pixelmonkey> Chipzz, the occasional internet cafe
[06:35] <pixelmonkey> Chipzz none of my own, anyway
[06:36] <Keybuk> zyga: err, could you elaborate what went wrong when, and what fixed it?
[06:36] <Keybuk> giftnudel: I'm going to go out on a limb and say it's something to do with not having sulogin atm
[06:37] <giftnudel> Keybuk: well it should be easy to reproduce, I will see if I can find something
[06:38] <Keybuk> cool
[06:38] <Keybuk> giftnudel: if you can reproduce it, I've something for you to try to see whether it fixes it
[06:38] <giftnudel> Keybuk: can I have the instructions now, then I will try it immediately
[06:41] <Keybuk> giftnudel: grab a copy of the sysvinit deb, extract the sulogin binary from it
[06:41] <Keybuk> and put that on the disk instead of my "exec /bin/sh" shell script that's hack-o-rama :p
[06:42] <pixelmonkey> dpkg-buildpackage is working great now, thanks for helping me through that one, guys.
[06:42] <zyga> Keybuk: after initial upstart from universe (the one needing a boot option) gdm/x crashed/were killed as soon as one logged in
[06:42] <zyga> then after a second gdm was running again and everything was okay
[06:42] <zyga> I tested this a couple of times without upgrading x or anything else
[06:42] <pixelmonkey> Oh, and good news.  My patch for gnome-power-manager works! :)
[06:42] <zyga> it was happening only with usptart as init
[06:43] <zyga> now, after upgrades (and amongs them the compat package) this issue is no longer present
[06:43] <giftnudel> Keybuk: ok, will tell you what happened
[06:44] <Keybuk> zyga: ahh!
[06:44] <Keybuk> zyga: yes, I know what that problem was ... X doesn't take kindly to flushing the console device under it
[06:44] <Keybuk> it was fixed in 0.2.0
[06:44] <zyga> ah :)
[06:44] <Keybuk> it was some code I found in sysvinit that I didn't really understand
[06:44] <zyga> good, it's a known and fixed issue then :)
[06:44] <zyga> heh :>
[06:44] <Keybuk> oddly enough, the same block exists in initng with /* what does this do? */ above it
[06:45] <zyga> :D
[06:45] <zyga> hehe
[06:45] <Keybuk> then I realised that sysvinit never calls it, at least not in any mode anyone's used since the 70s
[06:45] <Keybuk> so I took it out again
[06:45] <zyga> should be labeled /* magic */
[06:45] <zyga> I hope it's well commented now
[06:45] <Keybuk> nah, it's deleted now :p
[06:45] <zyga> hmm?
[06:45] <Keybuk> I could put it in a make_x_crash () function
[06:45] <Keybuk> but I don't see the need <g>
[06:45] <Amaranth> zyga: the code from sysvinit was crashing X
[06:46] <zyga> Amaranth: ahhhhhh
[06:46] <Amaranth> zyga: he removed it, X doesn't die
[06:46] <zyga> I thought that not keeping the code made x crash
[06:46] <zyga> OTOH it's nice to see such process
[06:46] <zyga> finding queer code that has been lurking for ages
[06:46] <zyga> that somehow controls remote stuff in unobvious manner
[06:47] <Keybuk> nah, was the code that "sets up and resets" a console device
[06:49] <pygi> Keybuk, congrats, nice work on upstart
[06:55] <ivoks_> right
[06:55] <ivoks_> great job Keybuk 
[06:56] <Keybuk> pygi, ivoks_: I'm reasonably pleased that it seems to be working well for most people that are trying it
[06:56] <giftnudel> Keybuk: Ok, the problem is different: As you pointed out, the console comes far too early. When fsck.ext3 is run, the problem is, that the console is there, when the check is run, and you don't see any progress of the check, the usplash just disappears and if you change to a different console, you can login while the check runs
[06:57] <pygi> ivoks_, I made silverspace try it yesterday :P
[06:57] <Keybuk> giftnudel: right, why did it reboot though?
[06:57] <giftnudel> Keybuk: putting sulogin in /sbin/ didn't change anything
[06:57] <giftnudel> Keybuk: what do you mean?
 Keybuk: I had xorg crash on me and after a reboot, I didn't have X there and logged into the console, I found out that e2fsk was running and apparently corrected some errors, and the system rebooted without a warning
[06:58] <Keybuk> "system rebooted without a warning"
[06:58] <giftnudel> I suspect fsck.ext3 rebooted 
[06:58] <pygi> Keybuk, will we have it as default already in edgy?
[06:58] <giftnudel> but since I didn't see anything, I don't know
[06:58] <Keybuk> pygi: I hope so
[06:58] <sladen> Keybuk: I've had crashed-X cause a reboot
[06:58] <Keybuk> giftnudel: it didn't reboot this time?
[06:59] <sladen> Keybuk: not with upstart mind
[06:59] <giftnudel> Keybuk: it did it again
[06:59] <ivoks_> uf, Keybuk i just saw i sent email to you, instead to the list; sorry
[07:00] <Keybuk> ivoks_: heh, nm, just send it to the list :)
[07:00] <giftnudel> Keybuk: It's surely the thing when it says "fsck corrected errors, system will reboot in 5 seconds", you just don't see it anywhere
[07:00] <Keybuk> giftnudel: oh, it does that?
[07:00] <giftnudel> yes
[07:00] <Keybuk> oh yes
[07:00] <Keybuk> heh
[07:00] <Keybuk> can you file a bug? :p
[07:00] <giftnudel> i will
[07:00] <Keybuk> thanks
[07:08] <giftnudel> Keybuk: bug 58609
[07:08] <Ubugtu> Malone bug 58609 in upstart "upstart doesn't display checkrootfs.sh messages" [Untriaged,Unconfirmed]  http://launchpad.net/bugs/58609
[07:10] <Keybuk> thanks
[07:36] <zyg1> ah, why is mvo never here on weekends :/
[07:37] <Keybuk> he has
[07:37] <Keybuk> a) an active social life
[07:37] <Keybuk> b) a lot of sports to play
[07:37] <Keybuk> c) a wife
[07:37] <Keybuk> :p
[07:40] <_ion> *shiver*
[07:40] <tseng> he doesn't even look old enough to be married
[07:40] <tseng> those germans
[07:41] <tseng> daniel and mvo look my age but are 5 years older
[07:41] <giftnudel> tseng: I still suspect that there are countries im which you marry earlier, in Germany, you need to be at least 18 ;)
[07:41] <zyg1> Keybuk: yeah I really understand
[07:42] <Keybuk> iirc mvo is older than me
[07:42] <tseng> goodness
[07:42] <zyg1> Keybuk: and I am married too you know :)
[07:43] <zyg1> I was rather saying that for people like me who can usually spend a small fraction of their time to cooperate the weekends are the only real time we have
[07:55] <imbrandon> Keybuk: do you know if Celso has had time to fix the dapper-backports thing , reason i ask is there is a pretty critical bug in the current amarok backport ( making it un-runable ) and it needs to be superceded
[07:55] <imbrandon> with the new one
[07:56] <Keybuk> I don't know either way, I'm afraid
[07:56] <Keybuk> I would assume not
[07:56] <imbrandon> k i figured not as it awas the weekend but i'm getting baraged with amarok people ;)
[08:05] <Seveas> Keybuk, can I pm you for a sec?
[08:30] <broonie> Is there anywhere I should be looking for Unbuntu-related init development work more generally than upstart? (Keybuk?)
[08:31] <Burgundavia> broonie: sorry, I don't folllow
[08:31] <Keybuk> broonie: in what kind of sense?
[08:32] <broonie> Issues to do with the boot process, especially things like ordering init scripts.
[08:32] <broonie> The sort of things that upstart is designed to facilitate.
[08:32] <Keybuk> upstart takes away those issues
[08:32] <broonie> It should end up pushing some changes into the rest of the world, shouldn't it?
[08:33] <broonie> Dependency information, for example.
[08:33] <Keybuk> upstart isn't dependency-based
[08:34] <Keybuk> let's start from the beginning
[08:34] <Keybuk> what is it that you're interested in, and what question are you trying to answer?
[08:34] <broonie> I'm trying to find out about any changes you guys are making to the init process.
[08:35] <broonie> My most practical interest comes from the NIS package.
[08:35] <Keybuk> previous changes in dapper (the re-ordering of the boot process) or future changes in edgy+1 (conversion to using upstart)?
[08:35] <broonie> Future ones.
[08:35] <Keybuk> upstart-devel is probably the best place for that, then
[08:36] <Keybuk> any specifications will be at least announced there
[08:36] <broonie> Right, OK. I'm already looking there so that's good.
[08:36] <Keybuk> there aren't any solid plans yet
[08:36] <broonie> "We've got some cool stuff, let's build on it." ATM?
[08:36] <Keybuk> yup
[08:36] <Keybuk> actually, we're "We've just finishing off the cool stuff, we'll build on it later" :)
[08:37] <broonie> If you don't build on it how will you get testing?
[08:37] <broonie> :P
[08:38] <Keybuk> for edgy I suspect we'll have the basic things like checkroot/mountroot/checkfs/mountfs/mountnfs/etc. converted to upstart jobs
[08:38] <Keybuk> as that's where our most annoying races are at the moment
[08:39] <love>  russian were are you
[08:40] <Keybuk> bear in mind that upstart jobs aren't init scripts
[08:40] <Keybuk> nor do they look like them
[08:40] <Keybuk> so we'll be happily contributing them to upstreams to include, if they wish
[08:41] <Keybuk> dunno whether they'll accept them or not, most don't ship init scripts because those are so wildly differently incompatible
[08:41] <zyg1> Keybuk: is there any list of packages that require upstart scripts and need love?
[08:42] <love>  
[08:42] <zyg1> love: not you, sorry :)
[08:42] <zyg1> I mean Ubuntu Love :D
[08:42] <Keybuk> zyg1: hmm, define "need" ?
[08:43] <zyg1> Keybuk: no one looks at providing upstart scripts for them
[08:43] <Seveas> upsart has jobs, not scripts 
[08:43] <zyg1> doh, right
[08:43] <zyg1> anyway..
[08:43] <zyg1> well they are scripts written using upstart job syntax :P
[08:44] <Keybuk> zyg1: still not sure I'm following you
[08:44] <zyg1> Keybuk: I was wondering if there is a list of packages that ship init.d script that could be modified to run from upstart but no one is looking at those packages because of low priority
[08:45] <broonie> Right. My most practical interest is in seeing if anyone manages to come up with a way of resolving the mess with things like NFS, NIS and autofs playing together. 
[08:45] <Keybuk> zyg1: well, for edgy, all packages are such
[08:45] <Keybuk> we're not going to start the mass-modification until edgy+1
[08:45] <Keybuk> broonie: this should, in theory, be easy with upstart
[08:45] <zyg1> so init stays by default? ok
[08:45] <broonie> Hence my interest in what you guys are up to.
[08:46] <Keybuk> zyg1: upstart will be /sbin/init by default
[08:46] <Keybuk> but upstart runs /etc/init.d/rc
[08:46] <Keybuk> so sysv-rc will be default
[08:46] <zyg1> Keybuk: I meant the 'old' init
[08:46] <zyg1> right
[08:46] <Seveas> zyg1, keybuk posted an implementation plan to -devel (or -devel-announce) recently
[08:47] <Keybuk> zyg1: /etc/init.d has absolutely nothing to do with sysvinit's init binary
[08:48] <Keybuk> it just happens that the configuration (/etc/inittab) causes it to be run
[08:48] <Keybuk> upstart can obviously be shipped with the same semantic configuration
[08:55] <Keybuk> the primary difference is that sysvinit will only start or stop processes (usually /etc/init.d/rc and getty) at a change of runlevel
[08:55] <Keybuk> where upstart will do it at any event
[09:01] <sladen> zyg1: fresh upstart pimpage at http://fridge.ubuntu.com/
[09:08] <zyg1> sladen: interesting, thanks
[09:11] <imbrandon> wow Keybuk nice^Wwell said blog post
[09:14] <zyg1> what's with all that debian bye bye stuff?
[09:15] <imbrandon> well its not nessesarly that that i was speaking of, more of the "why does debian single out ubuntu" when there is knappix,xandros,linspire and TONS of other dirivtives, i mean ubuntu clearly has a diffrent direction
[09:15] <imbrandon> not so much the "bye bye debian"
[09:16] <sladen> imbrandon: Ubuntu puts nitro in the gas tank
[09:16] <imbrandon> ;)
[09:17] <imbrandon> hahah that just gave me an idea to put a ubuntu/NOS mixture logo on my website ;)
[09:17] <zyg1> how about throwing tar files out the window and redesign internal deb format a little bit?
[09:17] <zyg1> (and dropping ar as well)
[09:18] <imbrandon> wow zyg1 were'd that come from
[09:18] <zyg1> imbrandon: my frustration with writing stuff that manages them :/
[09:18] <imbrandon> ;)
[09:19] <zyg1> apparently python's implementation of tar (tarfile module) has an annoying bug that makes it impossible to extract a single file from memory when the tar file is compressed 
[09:20] <zyg1> it has been wrongly labeled as win-specific and left ignored; it can be easily avoided by simply extracting a file to fs and reading it but the sheer crappines and complexity/legacy of tar is annoying IMO
[09:20] <sladen> zyg1: interesting, I haven't tried that.  only compressed from disk and uncompressed form memory
[09:20] <zyg1> sladen: w87
[09:20] <sladen> zyg1: perhaps you can feed the file through a stringio object
[09:20] <zyg1> wait
[09:20] <imbrandon> well honestly it does surprise me , i guesss since i came froma  non-debian background to ubuntu, but imho i do "the right thing" in that any "new" packages i get into debian first ( or try to ) and submit my patches upstream against their source etc etcetc , anyhow this isnt the channel for this so i'll shush now
[09:21] <zyg1> tarfile.open(mode="r|gz", fileobj=file("control.tar.gz")).extractfile("control").read()
[09:21] <zyg1> that's your crasher /
[09:21] <zyg1> do not take my suggestions seriously
[09:22] <zyg1> but I think that refreshing the design of debian packages is a valid goal
[09:24] <zyg1> a few key issues: random access, improved compression, simple access to metadata, dropping ancient tech
[09:25] <sladen> zyg1: we can actually get most of those using brain rather than redesign;  and we're working on it for delta-updates
[09:26] <zyg1> sladen: how can I use my brain to get random access in bz2 compressed archive?
[09:29] <zyg1> sladen: I just noticed your comment above stringio
[09:29] <zyg1> I already do at another level (after reading them from ar archive)
[09:30] <zyg1> I'll try doing that on control.tar.gz but that really equals with extracting the whole stuff into memory which is bad
[09:32] <Seveas> zyg1, why not use python_apt?
[09:32] <sladen> zyg1: scan the bz2 stream and note down a pointer to the corresponding start of each 900kB compressed block.  The you have ram-access with a sector size of 900kB
[09:32] <sladen> s/The/Then/
[09:32] <sladen> s/ram-access/random-access/
[09:34] <zyg1> Seveas: actually I'm trying to make it faster, I have a working version with apt_pkg and apt_inst
[09:34] <Seveas> aha
[09:34] <zyg1> but it is pretty slow (I guess it extracts everything in a way)
[09:35] <zyg1> sladen: I don't know the details of bz2 compression, do I need to read the entire stream to start decoding?
[09:36] <zyg1> (what I really want is to read and decompress just a header of some sort and the portions that actually contain the file I'm interested in)
[09:36] <sivang> zyg1: what are you developing?
[09:36] <zyg1> sivang: a package scanner for cnf
[09:36] <zyg1> sivang: I already got it working but it's quite slow
[09:36] <sivang> zyg1: nice, cnf being Linspire's installation from web mechanism ?
[09:37] <zyg1> ?
[09:37] <zyg1> no no
[09:37] <sladen> zyg1: no.  each 900kB in a bz2 stream is compressed completely independently.  You do need to know that the block for 3.5MB corresponds to offset 423,877 in the compressed stream though
[09:37] <zyg1> cnf => command not found, cnr is linspire tech called click and run
[09:37] <darius_> How can I see bugs only related to Edgy?
[09:38] <sladen> darius_: bugs that haven't been marked fixed are presumed to be present in edgy
[09:38] <zyg1> sladen: but that's really encoded in tar, right? I'd need to parse the tar header (I don't know if there is full map of stored stuff at the beginning) and the seek in bz2 to the place I want and start decoding there
[09:38] <darius_> sladen: thanks
[09:38] <zyg1> sladen: and since tar allows stuff to be appended I don't really think that's possible without reading the whole thing anyway
[09:38] <zyg1> I might be wrong but that's my gut feeling
[09:39] <sladen> zyg1: basically you need to have scanned the file once, and generated that "full map".  We intend to do this as standard for delta updates
[09:39] <zyg1> sladen: nice
[09:40] <zyg1> sladen: delta updates for stuff inside the archive with file granularity?
[09:40] <zyg1> or with block granularity without file correspondence?
[09:40] <zyg1> (or however that is spelled)
[09:41] <sladen> zyg1: so in this case we scan the file when it is generated on the build-daemon;  store it with the .deb and then partial HTTP gets you the random access to each compressed block
[09:41] <sladen> zyg1: no, sub-file granularity.  eg. to ~4kB
[09:41] <zyg1> sladen: I was thinking about something simple and backwards compatible
[09:42] <sladen> zyg1: this is backwards compatible
[09:42] <zyg1> ah, I was not thinking about delta updates
[09:42] <zyg1> about fast access
[09:42] <zyg1> deb is just an ar right?
[09:42] <sladen> zyg1: deb = ar -> gz -> tar 
[09:42] <zyg1> there are 3 files inside, debian-binary, control.tar.gz and data.tar.gz or bz2
[09:42] <zyg1> right
[09:43] <sladen> zyg1: so ignore the tar and ar parts.   the data you want is a gzip stream starting at an offset from the start of the .deb
[09:43] <sladen> zyg1: and then and a offset within that gzip stream
[09:43] <sladen> an
[09:43] <zyg1> sladen: parsing ar is easy, I already did that
[09:44] <zyg1> parsing tars and gzips is also easy - we have libs
[09:45] <zyg1> my only grief is that I need to fix one of them (tarfile) and do some stuff I don't want :)
[09:46] <zyg1> the thought of implementing tar again in python so that my use case is fast is not a nice thought though :/
[09:46] <Mithrandir> zyg1: just fix the tarfile module, then?
[09:46] <zyg1> right
[09:46] <zyg1> I'm doing that
[09:47] <sladen> zyg1: oh, and the python fuse module can't create device nodes.  That needs fixing too.  Upstream didn't so maybe I'll fix it straight in the Ubuntu package
[09:47] <sladen> ...it needs the API extending by one parameter (which are least is possible in pythong)
[09:48] <sladen> zyg1: so what are you working on;  it sounds like the core requirement might be very similar to the stuff that's already done
[09:48] <zyg1> sladen: yeah I know about mknod issue I used to play with fuse some time ago
[09:49] <zyg1> sladen: hmm?
[09:50] <zyg1> I only want to scan all debs I have and get symlinks, alternatives and executable files
[09:50] <sladen> zyg1: don't tell me you're writing debfs too :)
[09:50] <zyg1> no
[09:51] <sladen> zyg1: do tell! do tell!
[09:51] <zyg1> would be the perfect waste of my time IMHO :D
[09:51] <zyg1> since I already wasted my time on trying to get fuse to guess alternatives
[09:51] <zyg1> sladen: that's way too complex for my needs 
[09:53] <zyg1> in the end I need to tell someone that vim is available in the yada-yada set of packages...
[09:54] <sladen> ah, I see.  I think.  
[09:54] <zyg1> sladen: that's dead easy when you ignore alternatives
[09:54] <zyg1> (which unfortunatly cover the most interesting packages)
[09:55] <sladen> zyg1: mmmm
[09:55] <sladen> zyg1: it would be a wonderful problem to solve though
[09:56] <zyg1> I even parse the postinst scripts to get the alternatives but some !@#!% maintainers use $i for the target for some reason (like when they want vim.foo and vim.bar processed in a loop)
[09:57] <zyg1> after a dumb scan of main (and noticing that .tar.bz2 is valid as data source too) I got about 1.4K packages with $i or $var as alternative
[09:57] <zyg1> I don't want to scan universe without getting a better scanner which I'm trying to make now
[09:58] <zyg1> that's why I wanted to use fuse some time ago, so that I could chroot into fuse-based virtual fs and then exec the postinst script
[09:58] <zyg1> but that would take ages and would not solve the problem really :/
[09:59] <sladen> zyg1: you could do normal chmod and just wrap  update-alternatives
[09:59] <sladen> zyg1: normal chroot
[09:59] <zyg1> yeah but postinst stuff tend to do weird things
[09:59] <sladen> zyg1: but it would be better if they didn't need executing at all
[09:59] <zyg1> right
[10:00] <zyg1> (I'd really like update-alternatives in a declarative form)
[10:00] <sladen> zyg1: perhaps an option would be to add that meta-data to the packages that you can't parse
[10:00] <zyg1> so that 99% of packages don't use installers after unpacking data
[10:00] <zyg1> sladen: the important packages will get an eye review
[10:00] <sladen> zyg1: 1.4k as in 1,400 packages in main using  for $i in ... ; do update-alternatives...
[10:01] <zyg1> sladen: yes, 1.4k
[10:01] <zyg1> wait
[10:01] <sladen> zyg1: given that that about how many packages there /are/ in main
[10:01] <zyg1> no sorry 
[10:01] <zyg1> my mistake: first off that's for all arches
[10:02] <zyg1> and second that's counting packages the scanner couldn't understand
[10:02] <zyg1> 729 packages in main use alternatives with $variable
[10:02] <zyg1> that's the state from yesterday
[10:02] <zyg1> scanning main took about 7 hours
[10:03] <zyg1> scanning universe will take about a day since it's quite larger
[10:03] <zyg1> I don't event want to look at all packages in universe that require manual checks
[10:09] <HiddenWolf> Does anyone have any figures on how much these things get downloaded?
[10:10] <zyg1> HiddenWolf: there is a stat somewhere
[10:10] <zyg1> this popped up when breezy got noticed heavily
[10:10] <zyg1> I cannot remember the address though
[10:11] <HiddenWolf> zyg1: ah, shame
[10:12] <zyg1> http://torrent.ubuntu.com:6969/
[10:12] <zyg1> ah :)
[10:12] <zyg1> 185Gigs for i386 nice
[10:14] <HiddenWolf> zyg1: that _can't_ be the total traffic
[10:14] <HiddenWolf> I've uploaded 57gig myself alone.
[10:14] <zyg1> totals are at the bottom
[10:15] <zyg1> 1.1TB
[10:15] <zyg1> HiddenWolf: what link do you have!?
[10:15] <HiddenWolf> zyg1: just the one you gave me. :)
[10:15] <zyg1> no, your internet link :)
[10:15] <sladen> HiddenWolf: I suppose that torrent.ubuntu.com could just be spending most of its time seeding
[10:16] <sladen> HiddenWolf: you only get 1.5 CDs / GB
[10:16] <HiddenWolf> zyga: knot2 desktop i386
[10:16] <HiddenWolf> sladen: true that
[10:17] <zyg1> HiddenWolf: what kind of n
[10:17] <zyg1> internet link do you have, that's quite a speed :)
[10:17] <HiddenWolf> zyg1: 10/10
[10:18] <zyg1> at home?
[10:18] <HiddenWolf> zyg1: I live in a student building, converted office building. ISP is in the same building. ;)
[10:18] <HiddenWolf> zyg1: yes, at home. :)
[10:18] <zyg1> ah :D
[10:18] <HiddenWolf> pulling 1100kb/s now.
[10:18] <HiddenWolf> up
[10:19] <shining> wow
[10:32] <jdong> wow, is upstart fast at spawning a getty at startup :)
[10:32] <zyga> jdong: yeah, upstart is quite faster at getting the whole thing to boot :)
[10:32] <jdong> it's impressive
[10:33] <zyga> I think it's a bug/misfeature about the getty though
[10:33] <zyga> It's spawned from event.d 
[10:33] <imbrandon> so with the current state of upstart in edgy i can just install it and reboot and all should work ?
[10:33] <zyga> as soon as the system manages to run upstart
[10:33] <zyga> not after stuff happens (mounting, networking)
[10:34] <zyga> imbrandon: yeah
[10:34] <jdong> imbrandon: yeah, as long as you can reboot :P
[10:34] <aigarius> is knot2 really, really out? there is no link to its page from http://www.ubuntu.com/testing/
[10:34] <jdong> imbrandon: you're gonna have to take a hard reset :)
[10:34] <zyga> imbrandon: right, reboot doesn't work
[10:35] <zyga> use shutdown -r or something, read readme.Debian
[10:35] <jdong> reboot -f
[10:35] <jdong> or alt+sysrq+b
[10:35] <imbrandon> reboot -f
[10:35] <jdong> come on, alt+sysrq+b is more dramatic
[10:35] <imbrandon> hrm ok i got to run to my son's bday party but i'll try it when i get home
[10:35] <jdong> sync once or twice or three or 20 times first :P
[10:36] <imbrandon> bbiab
[10:36] <Mithrandir> jdong: and umount first.
[10:36] <Mithrandir> alt-gr+[sub] 
[10:36] <sladen> ?
[10:37] <jdong> Mithrandir: oh yeah, sysrq can remount-ro
[10:37] <jdong> forgot about that
[10:37] <jdong> meh, my reiser takes beatings from me quite well :-/
[10:37] <sladen> mmm reiserfs.  mmm no.
[10:37] <jdong> lol
[10:37] <jdong> it's not terrible nowadays
[10:38] <jdong> just don't expect reiserfsck to do much when hardware failures cause corruption
[10:38] <Mithrandir> hope you have backups, then.
[10:38] <jdong> Mithrandir: I am a paranoid backup freak
[10:38] <jdong> Mithrandir: I have at least 3 rsync'ed copies of my partitions lying around
[10:38] <sladen> even better;  keep those backups as reiserfs images *inside* a reiserfs partition
[10:39] <jdong> sladen: uhh, that's... not gonna make the reiserfsck game any more fun :)
[10:39] <jdong> so, now with upstart, how should I use sysv init scripts?
[10:40] <zyga> jdong: as you did before apparently
[10:40] <jdong> zyga: really?
[10:40] <zyga> yeah
[10:40] <jdong> I thought there was some initctl thing I should use instaed?
[10:40] <sladen> jdong: as normal for the moment
[10:40] <jdong> cool
[10:40] <sladen> jdong: at the moment, upstart is just starting /etc/init.d/rc
[10:48] <jdong> grr, squid and apt really don't play nice
[10:52] <jdong> mono-xsp* are uninstallable under edgy :-/
[10:54] <cbx33> hey pygi 
[10:54] <cbx33> know much about gconf?
[10:56] <pygi> cbx33, I know nothing ^_^
[10:56] <cbx33> hehe
[10:56] <cbx33> ok
[10:56] <zyga> cbx33: what do you want to know?
[10:57] <cbx33> um...about specifying sources
[10:57] <cbx33> xml:merged:path
[10:57] <cbx33> type things
[10:57] <zyga> hmm, sorry I don't know much about that
[11:38] <jelmer> infinity: ping