[12:04] <lupus_> is there a spec from freedesktop.org
[12:04] <daniels> what sort of psec?
[12:05] <lupus_> where to save favourites, my documents, my music
[12:05] <spotter> how does one change the ubuntu locale, so that I get color instead of colour
[12:05] <daniels> lupus_: no
[12:06] <lupus_> I was thinking of using .Documents/ and let gnome-vfs for example create my documents or in my language mijn documten 
[12:06] <lupus_> to it
[12:07] <lupus_> so you can store somewhere and the name of the location is translatable since it is handled by gnome-vfs
[12:09] <jdub> lupus_: the design proposed by alex will mean translated names in the gui, normal names on the filesystem
[12:09] <jdub> not dot directories, either
[12:09] <jdub> this is the same way os x does it
[12:10] <lupus_> where is this method discribed?
[12:10] <jdub> o don't think there's a single document
[12:10] <jdub> s/^o/i
[12:11] <lupus_> couldn't that design be used for downloads instead of desktop?
[12:13] <jdub> no, that's a different problem
[12:14] <jdub> the downloads stuff is a matter of choosing a good location
[12:14] <jdub> the translation stuff is just a way of seeing what's on disk differently
[12:30] <robertj> jdub: is synaptic coming out of the Computer menu when gnome-app-install gets done?
[12:31] <jdub> robertj: most likely
[12:38] <lupus_> jdub, http://www.osnews.com/story.php?news_id=8888 I think this will interest you ( Reducing Fedora Boot Time )
[12:42] <jdub> lupus_: mmm, posted links to ubuntu-devel previously :)
[12:44] <lupus_> :) I hadn't openened the png's so I thought those where new :)
[12:50] <amu> n8 
[12:58] <Keybuk>  * Creating copy of debian gaim
[12:58] <Keybuk>  * Considering public_html/ongoing-merge/gaim/gaim_debian.patch
[12:58] <Keybuk>    - 1 patch hunks dropped
[12:58] <Keybuk>  * Creating copy of ubuntu gaim
[12:58] <Keybuk>  * Considering public_html/ongoing-merge/gaim/gaim_ubuntu.patch
[12:58] <Keybuk>    - All patch hunks applied
[12:58] <Keybuk>  * Using result of applying ubuntu patch
[12:58] <Keybuk> \o/
[12:58] <Keybuk> I found one
[12:58] <Keybuk> (invert the "result of applying X patch" bit -- it's the wrong way round)
[12:58] <daniels> nice
[12:59] <Keybuk> actually, sorry, it's the "Creating copy of X gaim" that's the wrong way round
[01:00] <Keybuk>  * Creating copy of ubuntu gaim
[01:00] <Keybuk>  * Considering public_html/ongoing-merge/gaim/gaim_debian.patch
[01:00] <Keybuk>    - 1 patch hunks dropped
[01:00] <Keybuk>  * Creating copy of debian gaim
[01:00] <Keybuk>  * Considering public_html/ongoing-merge/gaim/gaim_ubuntu.patch
[01:00] <Keybuk>    - All patch hunks applied
[01:00] <Keybuk>  * Using result of applying ubuntu patch
[01:00] <Keybuk> ... there, fixed
[01:01] <jdub> whee :)
[01:03] <thom> man, rendezvous in gaim is SO FUCKED
[01:07] <Kamion> ... and watches it crash and burn, oh dear
[01:07] <daniels> Kamion: nice!
[01:07] <daniels> heh
[01:15] <Keybuk> Kamion: work directories should be gone too <g>
[01:16] <Safari_Al> jdub, thanks for that email.
[01:18] <robertj> Is partition resizing on the list for hoary?
[01:23] <Kamion> what kind of partition resizing?
[01:26] <lamont_r> Nov 17 00:25:41 buildd: Nothing to do -- sleeping 300 seconds
[01:26] <lamont_r> woot
[01:29] <lamont_r> now I'm stuck with fixing cyrus-sasl2 sometime.
[01:29] <lupus_> ntfsresize is in sid installer, will also be in ubuntu?
[01:29] <Keybuk> lamont_r: want the new set of merges all assigned to you?  if you're bored? :p
[01:29] <lamont_r> how big is the set?
[01:30] <Keybuk> 20 today
[01:31] <Keybuk> (this is actually about 5 days worth ... because some idiot commented out the "download the new needs-merged.txt file" line)
[01:34] <lamont_r> hehe
[01:34] <lamont_r> you could certainly assign them to me...
[01:37] <lamont_r> I'd eventually forgive you...
[01:40] <Kamion> lupus_: already is in hoary
[01:40] <robertj> Kamion: sorry, ahead of myself. On-the-fly NTFS resizing
[01:41] <Kamion> robertj: see what I said to lupus_ then
[01:41] <Kamion> robertj: download Array CD 1, try it out
[01:41] <robertj> Array?
[01:42] <Kamion> see ubuntu-users mailing list archives ...
[01:42] <Kamion> Subject: Array CD 1
[01:42] <Keybuk> https://bugzilla.ubuntu.com/buglist.cgi?bug_status=NEW&bug_severity=normal&field0-0-0=alias&type0-0-0=regexp&value0-0-0=%5Emerge-
[01:42] <Keybuk> ^ that's the list of new, unassigned "main" merges
[01:42] <robertj> ahh
[01:44] <Kamion> garrr, this installer crash is really hard to reproduce under any kind of controlled conditions
[01:46] <lupus_> Kamion, I'm supprised ubuntu uses ntfs since redhat doesn't dare to touch it because of IP problems
[01:47] <Kamion> lupus_: reference for that claim?
[01:47] <Kamion> lupus_: we don't use it for anything important, it's just the stuff from the stock kernel
[01:47] <lupus_> I know
[01:47] <lupus_> but redhat does not include the ntfs kernel module
[01:49] <lupus_> Why doesn't my Distro support NTFS?
[01:49] <lupus_> Due to the uncertain legal status of using the NTFS driver, RedHat and Fedora have chosen to leave the driver out of their kernels. 
[01:49] <lupus_> http://linux-ntfs.sourceforge.net/rpm/about.html
[01:49] <Kamion> it's not something that's come up, and it doesn't seem to have bothered anyone in Debian either
[01:50] <lupus_> so it ntfs is not only in non_US or something in debian?
[01:51] <Kamion> it's in the Debian stock kernels; non-US is essentially dead
[01:51] <lupus_> ic :)
[01:53] <lamont_r> elmo: I see you arrived safely
[01:54] <Keybuk> the George does seem to have utterly replaced the Claverley as the Canonical hotel now
[01:55] <jdub> thom: it should be fixed up to use howl
[01:57] <jdub> Keybuk: is it better?
[01:57] <lamont_r> is the George better?
[01:57] <thom> jdub: yes
[01:58] <thom> there's a launchpad sprint and a baz sprint?
[01:58] <Keybuk> I've not actually sampled the George yet
[02:00] <elmo> the George is so much better it's not funny
[02:00] <elmo> decent 2Mb internet, okay rooms, good food
[02:00] <Keybuk> but no cute staff :'(
[02:01] <elmo> dude, you just said you haven't been here yet?
[02:01] <jdub> elmo: ahr, that's reasonable
[02:01] <Keybuk> elmo: I asked others who rated them for me :)
[02:02] <elmo> oh, yes, now I remember kinnison complaining about the porter who bought up his special pillow :)
[02:05] <Keybuk> his "special" pillow :)
[02:07] <Kamion> do I want to know?
[02:07] <jdub> Kamion: waterproof.
[02:08] <Keybuk> it's the "special" sheeting you need to worry about
[02:08] <jdub> lamont_r: looked into the postfix disk access stuff?
[02:09] <lamont_r> huh?
[02:09] <lamont_r> vs laptop you mean?
[02:10] <lamont_r> jdub?
[02:14] <Kamion> aaargh
[02:14] <Kamion> running /sbin/debian-installer under strace is sufficient to make the problems go away
[02:16] <Kamion> jdub: could you add an xserver-xorg package to bugzilla please?
[02:16] <robertj> Kamion: you may need to proceed onto more intense debugging techniques such as banging hand in cabinet or, for the most challenging bugs, slamming nuts into car-door
[02:25] <lifeless> robertj: thats vicious. 
[02:25] <lifeless> OW.
[02:48] <robertj> Kamion: do you use bochs to debug d-i?
[02:48] <Kamion> robertj: no, I have plenty of real test hardware which is much faster
[02:48] <robertj> netboot?
[02:49] <Kamion> usually CD since that's the primary Ubuntu installation method
[02:49] <Kamion> occasionally netboot
[02:49] <robertj> hehe, how many times does a rewriter rewrite ;)
[02:49] <Kamion> not had enough coasters yet to worry about it
[02:50] <Kamion> maybe one or two out of I suppose many hundreds of writes
[02:51] <Kamion> the main disadvantage of using CDs is that they're slow
[02:53] <robertj> is there a way to take screenshots during the install?
[02:54] <lupus_> can someone look if he also has this in /etc : drwxr-s---   2 root   dip         96 2004-11-15 22:37 chatscripts
[02:55] <Kamion> robertj: not without some kind of emulation
[03:32] <Kamion> erk, I think I'm hitting some kind of library reduction bug
[03:32] <Kamion> damnit
[03:33] <Kamion> adding /usr/bin/strace to the image is sufficient to make it work
[03:57] <jdub> lamont: yeah
[03:57] <jdub> Kamion: sure
[04:01] <jdub> Kamion: done
[04:19] <jdub> what's this mondo/mindi stuff?
[05:07] <shaya> I think I discovered a regression in ubuntu's gnome over plain debian
[05:07] <shaya> in plain debian, if I click on the dvd device (with a dvd movie inside) it auto launches totem
[05:08] <shaya> in ubuntu, gnome-device-manager takes care of the initial load of totem (which doesn't work in debian) but every other open action, just opens the dvd in nautilus
[06:43] <infinity> shaya : Is that a 2.6 versus 2.8 thing, or a Debian versus Ubuntu thing?
[07:06] <fabbione> morning guys
[07:18] <infinity> You've had reports via IRC?
[07:18] <infinity> On non-UML hosts?
[07:18] <infinity> (That .mem file is sparse on every non-UML host I've run across, and never large enough to be an issue)
[07:19] <infinity> fabbione : Also, this is clearly the wrong channel.  Pretend I was typing in #debian-apache.
[07:19] <infinity> fabbione : I'm half asleep. :)
[07:19] <lifeless> mjg59: are there plans to get your kernel into hoary? (I'd love to have a prop ati driver package :))
[07:35] <fabbione> infinity: don't worry :-), but yes.. i got the same report on a non uml host
[07:41] <infinity> fabbione : Was the report "I ran out of space on /tmp" or was it "there's a giant file in /tmp and it scares me!"?
[07:41] <infinity> fabbione : Cause I've had a lot of the latter before (which goes away when you tell them "it's a spare file, retard")
[07:41] <infinity> s/spare/sparse/
[07:42] <infinity> fabbione : Note that until very recently, PHP/libmm on unstable worked this very same way.
[07:42] <fabbione> infinity: apache was crashing on restart. 
[07:42] <fabbione> downgrade mm was fine again
[07:42] <fabbione> same setup except outside UML
[07:42] <infinity> Must have been a tiny /tmp
[07:43] <infinity> Or stable's libmm is broken in even more interesting ways than unstable's.
[07:43] <fabbione> also.. if it was working before with a 1MB /tmp it must keep working if nothing else changed
.. I agree.  The change should have been tested more.
[07:43] <fabbione> infinity: libmm in stable got an update in the "attempt" to fix a php segfault
[07:44] <infinity> Not much we can do about it now, unless Joey can push a fix in somehow.
[07:44] <fabbione> infinity: i am going to knock on Joey's door if he doesn't fix it now
[07:44] <fabbione> or i will let him adopt apache... so he can feel the pain
[07:45] <infinity> (FOr the reocrd, I had no idea broonie had uploaded that s-p-u upload until it made it into the last update... I only expected him to do the MMFILE experimenting on unstable..)
[07:45] <fabbione> read the changelog...
[07:45] <infinity> And, for added irony, I've since completely removed MM as a session handler in PHP, cause it sucks.
[07:45] <fabbione> it is scary
[07:45] <fabbione> brb
[08:56] <fabbione> hey lamont
[09:14] <lamont_r> ho
[09:14] <lamont_r> yo, even
[09:14] <fabbione> lamont_r: i saw all the build failures for x.org on ia64 :-)
[09:14] <lamont_r> heh
[09:14] <lamont_r> how many?
[09:14] <fabbione> if you can send me the files i asked for, i can fix it before we upload the next version
[09:15] <fabbione> several failures...
[09:15] <fabbione> because it's missing rman :-)
[09:15] <fabbione> an orgy of failures :P
[09:15] <lamont_r> yeah - it's still building from a partial archive
[09:15] <fabbione> ok
[09:16] <lamont_r> and then I'll have to really fix cyrus-sasl2 tomorrow...
[09:16] <fabbione> and it is trying at a regular rate of once every hour :-P
[09:16] <lamont_r> of course, nearly time to fall over tonight.
[09:16] <lamont_r> really... hrm. yeah.
[09:16] <fabbione> did you spot the problem with my logs?
[09:16] <lamont_r> not yet
[09:17] <fabbione> ok
[09:17] <lamont_r> more to the point, I didn't spot how your's managed to avoid the failure...
[09:17] <fabbione> because my sparc is more cool than your buildd's?
[09:17] <lamont_r> no
[09:17] <lamont_r> :-)
[09:17] <fabbione> hheeh
[09:18] <fabbione> interesting
[09:18] <fabbione> take a look at zsh
[09:18] <fabbione> did you install autoconf on amd64 and ia64 buildd manually?
[09:19] <fabbione> because i get the same FTBFS as i386 and ppc
[09:20] <lamont_r> no
[09:20] <lamont_r> $alternatives{"automaken"}="automake1.8";
[09:20] <lamont_r> $
[09:20] <fabbione> possible leftovers?
[09:20] <lamont_r> I do have that...
[09:20] <fabbione> which file?
[09:21] <lamont_r> .sbuildrc
[09:21] <lamont_r> I suppose it's possible that autoconf got installed and left somehow, but I don't believe so
[09:22] <fabbione>         "xserver"                       => "xserver-xfree86",
[09:22] <fabbione> hmm
[09:22] <fabbione> there are a bunch of those that needs love
[09:22] <infinity> Nothing build-deps on xserver anymore.
[09:22] <lamont_r> yeah - mail-transport-agent is another
[09:22] <lamont_r> which is good..
[09:23] <infinity> Rather, yes.
[09:25] <fabbione> infinity: a lot.. and you will fail on X.org :-)
[09:25] <fabbione> i still need to merge the MANIFEST files check
[09:25] <fabbione> but the code compiles without any glitch
[09:25] <seb128> morning
[09:25] <fabbione> morning seb128 
[09:25] <infinity> fabbione : How many regressions have you guys found in moving from XF86 to Xorg?
[09:26] <seb128> hi fabbione 
[09:26] <fabbione> infinity: from a usage pov or packaging pov?
[09:27] <infinity> fabbione : Usage.
[09:27] <fabbione> infinity: only 3 or 4, but nothing really bad
[09:27] <fabbione> we have only 3 bugs open on X.org
[09:27] <fabbione> one of them fixed upstream
[09:27] <infinity> You'll get a few more if you push it into Debian, I'd imagine. :0
[09:28] <fabbione> infinity: i am sure about that too, but we need to merge first
[09:28] <fabbione> and Overfiend is still drawing the layout for the SVN repo
[09:28] <fabbione> after that we can start syncing again
[09:28] <infinity> No chance of a hostile takeover, I guess?
[09:29] <fabbione> why a hostile takeover?
[09:29] <fabbione> Overfiend and I work perfectly together
[09:29] <fabbione> there is NO point in taking over
[09:29] <infinity> Fair enough.  But the lack of DanielS in the team seems like a mistake.
[09:30] <infinity> Clash of egos notwithstanding.
[09:30] <seb128> fabbione: how is the UTF-8 issue going ?
[09:30] <fabbione> infinity: i have been pushing changes back and forward in place of Overfiend and DanielS
[09:30] <infinity> Heh.
[09:30] <fabbione> infinity: so that is not an issue
[09:30] <fabbione> seb128: it's not.. it is not high priority for me at the moment
[09:31] <seb128> ok
[09:31] <fabbione> seb128: you will have to bug daniels at least until DecConf
[09:31] <seb128> fabbione: there is no hurry, I was just wondering
[09:31] <fabbione> seb128: i know :-) i don't feel the pressure
[09:37] <fabbione> night lamont
[09:37] <seb128> 'night lamont_r 
[09:52] <fabbione> sabdfl morning 
[09:53] <sabdfl> hiya fabbione
[10:07] <seb128> daniels: here ?
[10:10] <daniels> seb128: yo, sup
[10:10] <seb128> daniels: 
[10:10] <seb128> xine-lib (1-rc6a-1ubuntu1) hoary; urgency=low
[10:10] <seb128>   * Resync with Debian.
[10:10] <seb128> that's not really useful to redo a merge with the new version
[10:10] <seb128> do you any details on what do you merged ?
[10:12] <daniels> i took the ubuntu changes ... and applied them to the debian version ... like the rest of the sync ...
[10:12] <seb128> ok, so I just have to read the diff to get the ubuntu changes ? :)
[10:12] <seb128> that's only the 
[10:12] <seb128>   * debian/control:
[10:12] <seb128>   - Also build-depend on libtheora-dev, because it is worthy of our praise and
[10:12] <seb128>     adulation. Also, tautologies.
[10:12] <seb128> 
[10:12] <seb128> ?
[10:13] <calc> whee!
[10:13] <calc> someone using theora finally
[10:13] <seb128> ah ah
[10:13] <daniels> seb128: um, yeah, dude
[10:14] <seb128> ok, thanks
[10:14] <calc> daniels: canonical hackfest in london?
[10:15] <infinity> Hotel with a dinky 2Mb pipe.  You could DoS it and make a great number of Canonical folk very unhappy.
[10:16] <infinity> Not that I'd suggest such a thing.
[10:16] <elmo__> infinity: all two of them
[10:16] <infinity> Oh, that's less fun, then.
[10:16] <infinity> And lulu...
[10:16] <infinity> That's 3.
[10:16] <elmo> infinity: dude, this IP isn't the hotel
[10:16] <infinity> Oh. :)
[10:17] <daniels> (the hotel is the one with 'hotel' in the domain name)
[10:17] <calc> george.kkhotels.co.uk
[10:18] <daniels> i'm glad I checked in as 'george rumpadinkle' :P
[10:20] <calc> heh
[10:21] <Mitario> hello
[11:10] <Astharot> hello
[11:37] <azeem> daf: did you repackage the ruby-gnome2 upstream tarball? The size seems to be slightly off to what's on upstream's download site, or I picked the wrong tarball
[01:05] <seb128> elmo: gaim sync please :)
[01:05] <seb128> jamesh: here ?
[01:11] <elmo> seb128: done
[01:11] <seb128> thanks
[01:12] <Kamion> jdub: ... with Fabio as maintainer? (xserver-xorg)
[01:13] <daniels> Kamion: hmm?
[01:17] <Kamion> daniels: unless there's a Bugzilla account that goes to both of you
[01:17] <daniels> Kamion: assigned to d.s@c.c please
[01:17] <Kamion> Fabio's the xserver-xfree86 maintainer at the moment; should it be you for xserver-xorg?
[01:17] <Kamion> ok
[01:17] <Kamion> jdub: ^--
[01:17] <daniels> Kamion: xorg is mine now, yah
[01:18] <fabbione> Kamion: *X* -> daniels
[01:22] <Kamion> bonus for Fabio, then :-)
[01:22] <daniels> heh, yeah
[01:29] <fabbione> sparcbuildd@vultus5:~/chroots$ sudo rm -rf xorg-chroot/
[01:29] <fabbione> daniels: sparc is done and changes committed to baz
[01:29] <daniels> awesome, will update it
[01:29] <daniels> nice work :)
[01:30] <fabbione> i need to find the time to merge m68k
[01:41] <stratus> Do i need report that sed 4.1.2-2 postinst is f*cked?
[01:41] <sid77> lol
[01:42] <sid77> just reported (again) in #ubuntu
[01:42] <stratus> hmm 3771 maybe
[01:42] <daniels> stratus: no
[01:42] <daniels> stratus: known problem, came in via a debian sync, will be fixed via a debian sync
[01:42] <sid77> btw, is there anything to do beside waiting?
[01:42] <sid77> ok
[01:43] <sid77> answered by daniels 
[01:43] <stratus> daniels, i see
[01:43] <Kamion> the sed thing has been reported about 10 times so far
[01:43] <stratus> sid77, bugzilla.ubuntu.com/3771
[01:43] <stratus> no problem i was just asking and searching at bugzilla at the same time
[01:43] <sid77> stratus, I know
[01:43] <daniels> Kamion: my My Bugs query now makes me look a hell of a lot worse
[01:44] <Kamion> it's already been fixed in Debian incoming
[01:44] <Kamion> daniels: heh
[01:44] <Kamion> daniels: try mine
[01:44] <stratus> yes, 4.1.2-3 is there
[01:47] <Kamion> daniels: (100 bugs or thereabouts; I made a heroic effort the other day to get it down from 120-ish)
[01:48] <daniels> Kamion: wow, nice
[01:58] <fabbione> daniels: -1 hour and 2 minutes
[01:59] <daniels> -1?
[01:59] <fabbione> countdown
[02:00] <fabbione> 1 hour to go
[02:00] <fabbione> now
[02:00] <fabbione> tic
[02:00] <fabbione> tac
[02:00] <fabbione> tic
[02:00] <fabbione> tac
[02:07] <amu> moins
[02:08] <haggai> amu!
[02:09] <amu> haggai: *waves*
[02:09] <seb128> elmo: python-gnome2 sync too :)
[02:12] <elmo> done
[02:13] <seb128> thanks
[02:16] <Kamion> elmo: please sync console-data
[02:16] <Kamion> (-46.1)
[02:17] <elmo> done
[02:17] <Kamion> thanks
[02:34] <Kamion> elmo: please sync redland from Debian
[02:36] <Kamion> boggle, now d-i/udev just works; I didn't do anything to fix it
[02:36] <Kamion> last night main-menu was segfaulting away
[02:41] <Mithrandir> hmm, are we going to move to gcc-3.4 for hoary?
[02:42] <elmo> Kamion: done
[02:43] <Kamion> thanks
[02:51] <fabbione> daniels: i think i found a really really bad bug in xserver-xorg
[02:51] <daniels> cool!
[02:51] <daniels> what is it?
[02:51] <daniels> (please don't say it's about xkb, or I'll cry)
[02:52] <fabbione> i need to see if i can reproduce it but it i think we lost debconf data on conversion from xserver-xfree86
[02:52] <daniels> gah
[02:52] <daniels> what data?
[02:53] <fabbione> at least 3 things... monitor identifier, and vert/horiz sync
[02:54] <daniels> mmm
[02:54] <fabbione> it happend on the hp box
[02:54] <fabbione> not on mine
[02:55] <daniels> so what's the effect of losing V/H?  recalculated from resolution?
[02:55] <fabbione> i need to try and reproduce it
[02:55] <fabbione> daniels: you are the X.org maintainer.. you need to tell me :P
[02:55] <fabbione> no seriously
[02:56] <fabbione> i need to check on a 3rd box
[02:56] <daniels> hey man, you know my plans for sync rates being written out anyway :)
[03:04] <daniels> fabbione: hey, here's a cool datapoint
[03:04] <daniels> when we use libX11 from the modular tree, UTF-8 locales work.
[03:05] <fabbione> daniels: good.. now diff them and find the fix :P
[03:06] <daniels> ha ha ha
[03:07] <fabbione> HMMM
[03:12] <fabbione> daniels: apparently it was a glitch on the hp
[03:12] <fabbione> the other 4 installations are ok
[03:12] <fabbione> daniels: please verify witrh debconf-show xserver-xorg
[03:12] <fabbione> if you get all the values as expected
[03:12] <fabbione> guys.. you too
[03:18] <daniels> fabbione: ok, the UTF-8 problem isn't related to the locale *data*, it's a code issue
[03:18] <daniels> (aieeeeee!0
[03:19] <daniels> fabbione: i have h-s and v-s
[03:22] <fabbione> daniels: ok good. it's a glitch on the hp box than
[03:22] <fabbione> i will have to check it later
[03:30] <michiel_> hi everyone
[03:31] <sivang> hi pitti
[03:31] <Mitario> *sigh* great
[03:36] <Mitario> i think the breakage of the sed package is not new? :)
[03:36] <Kamion> Mitario: new today in Debian, also fixed in Debian incoming
[03:36] <Mitario> ah ok
[03:36] <Kamion> reported lots :)
[03:36] <Mitario> yeah, thought so :)
[03:36] <seb128> the package which is in incoming doesn't build here :/
[03:43] <elmo> Kamion: rock on
[03:44] <daniels> fabbione: found the spots where threading is TOTALLY BROKEN
[03:44] <fabbione> daniels: good
[04:01] <Kamion> seb128: built fine on the Debian buildds
[04:02] <elmo> Kamion: did you hear back from Lamont WRT Herbert and ia64 ?
[04:02] <Kamion> elmo: no
[04:03] <elmo> ok, I'll mail him - I need some kernel love for ia64 myself anyways
[04:18] <Kamion> elmo: we *are* still contracting Herbert, aren't we? haven't heard anything from him in a while
[04:19] <seb128> Kamion: apparently the problem is locale dependent
[04:19] <elmo> Kamion: yep, I talked to him about kernel  security stuff yesterday
[04:19] <bob2> selinux!
[04:19] <elmo> bob2: is that some variation on tourettes?
[04:21] <Kamion> elmo: ah, good
[04:21] <Kamion> seb128: yay :-/
[04:21] <Kamion> elmo: we should have a swear-box for buzzwords
[04:21] <daniels> Kamion: XORG COMPOSITE SHINY
[04:22] <Kamion> *bzzt* *bzzt* *bzzt*
[04:22] <Kamion> (udev!)
[04:22] <daniels> BING!
[04:30] <spotter> is there any reason universe packages don't appear in the gnome menu, even when they have a .desktop file (and appear in plain debian)
[04:30] <spotter> for example, blam
[04:31] <rburton> spotter: tried killing the panel?
[04:31] <spotter> argh
[04:32] <rburton> that will force the panel to re-read
[04:32] <spotter> it usually does it by itself w/ot that
[04:33] <rburton> yes
[04:33] <rburton> this is a bug, do you have fam/gamin installed?
[04:34] <spotter> gamin is installed
[04:34] <spotter> and so is fam
[04:34] <spotter> hmm
[04:34] <spotter> that cant be a good thing
[04:35] <spotter> hmm, guess they dont conflict
[04:35] <spotter> thought they did
[04:37] <spotter> hmm, dpkg seems to allow one to remove dependencies
[04:59] <azeem> hoary features evolution-2.1, right?
[04:59] <Keybuk> yes
[05:00] <azeem> thanks
[05:02] <pitti> sjoerd: just read the hal commit changelogs; looks nice :-)
[05:03] <sjoerd> pitti: thanks
[05:03] <sjoerd> pitti: one issue though.. My external usb2 disk doesn't have the sysfs removable property set to 1
[05:04] <pitti> sjoerd: now that you tried out tarball.mk, do you actually like it? You should not use it just because I do :-)
[05:04] <pitti> sjoerd: bad
[05:04] <sjoerd> dbs-edit-patch is indeed very nice
[05:04] <pitti> sjoerd: but pmount does not only rely on this attribute
[05:04] <pitti> sjoerd: it uses the attribute in addition to bus checking
[05:04] <sjoerd> but to set it up correctly is a hassle 
[05:05] <sjoerd> pitti: that doesn't help hal :(
[05:05] <pitti> sjoerd: all USB and FireWire devices should work (also with older kernels, which don't provide the attribute anyway)
[05:05] <pitti> sjoerd: ah, I see what you mean
[05:05] <pitti> sjoerd: does it work with other devices?
[05:05] <sjoerd> yah my usb flash pen thingy works fine
[05:06] <sjoerd> brw-r-----  1 root hal 8, 33 Nov 16 22:44 /dev/sdc1
[05:07] <sjoerd> pitti: currently hacking gvm so we can just specify "/usr/bin/pmount-hal %h"
[05:07] <pitti> oh, you changed the permissions to 0640
[05:07] <pitti> sjoerd: but I already modified g-v-m to do that
[05:08] <pitti> sjoerd: ah, you mean you support the %foo templates?
[05:08] <sjoerd> yeah
[05:08] <pitti> like %d for device, %u for udi and so on?
[05:08] <sjoerd> yes
[05:09] <elmo> pitti: 'Using sudo is inherently insecure' - what crack are you on, and can I have some please?
[05:10] <daniels> elmo: please dude, no more :P
[05:10] <pitti> elmo: because as soon as the user sudo'ed something, all of his processes have essentially root capabilities
[05:10] <Keybuk> pitti: that's the same as "su" though
[05:11] <pitti> right
[05:11] <Keybuk> "user who can become root" in "can become root" shocker
[05:11] <pitti> well, only if you have the timeout, which su does not have
[05:11] <Keybuk> *shrug* just send key presses to the terminal with su running inside it
[05:11] <Keybuk> it's not exactly rocket science
[05:11] <pitti> right
[05:12] <Keybuk> sudo encourages you to *not* leave a root shell open all the time
[05:12] <pitti> sudo is just a little worse, but it does not open principially new holes
[05:12] <Keybuk> I'd say sudo is a little better
[05:12] <pitti> Keybuk: but you can gain root even if you don't have a root shell open
[05:12] <Keybuk> only within a few minutes of the user becoming root themselves
[05:12] <elmo> pitti: dude, there is more to sudo than "unrestricted root"
[05:12] <pitti> Keybuk: I think sudo and su don't differ quite much in this regard
[05:12] <pitti> elmo: right
[05:13] <pitti> elmo: but in Warty it's unrestricted root
[05:13] <pitti> elmo: that's why I said that I don't bother to fix the bug in Warty
[05:13] <pitti> elmo: it can be fixed in Hoary and upstream, for my sake
[05:13] <pitti> elmo: well, if there is a general desire, I do fix the bug in Warty
[05:13] <pitti> but I don't really see the benefit
[05:14] <Keybuk> ANYTHING that lets you elevate your privileges is an attack honeypot
[05:14] <pitti> fix a small hole and neglect the really big one? hmm
[05:14] <elmo> pitti: no, dude it's not "unrestricted root" in warty
[05:14] <elmo> I have 40 servers that say otherwise
[05:14] <pitti> logging in as root at a text console is the only safe way anyway
[05:14] <Keybuk> there is absolutely no way to allow someone to "become an administrator" which isn't a gaping security hole
[05:14] <elmo> please don't make inane assumptions like users never customize 
[05:16] <pitti> elmo: do your servers call sudo programs without specifying the exact path?
[05:16] <pitti> elmo: because the attack only works if you call programs without path
[05:16] <pitti> but then you can as well put a trojan horse somewhere else in $PATH
[05:16] <pitti> Well, for my sake I fix it
[05:17] <elmo> no, they don't - what's that got to do with anything tho?    my point is that the default, is just that, the default.  it's not the only way sudo is used
[05:17] <pitti> Keybuk: what's wrong with login at the text console?
[05:17] <daniels> pitti: the fact it's at a data centre?
[05:18] <pitti> daniels: well, root login over ssh should be safe too, right? The pty belongs to root
[05:18] <Keybuk> pitti: uh, no
[05:18] <Keybuk> the terminal is owned by the USER
[05:18] <Keybuk> simply debug the terminal and send key presses to it
[05:18] <Keybuk> and that's not even considering anything more evil
[05:19] <pitti> Keybuk: no, it's not; if I login as ROOT, the terminal is owned by root
[05:19] <Keybuk> pitti: login as root where?
[05:19] <Keybuk> the machine's in the data centre
[05:19] <pitti> Keybuk: text console?
[05:19] <pitti> Keybuk: or ssh
[05:19] <daniels> pitti: send key events via XTest
[05:19] <pitti> daniels: I speak about text consoles, not X terminals
[05:19] <Keybuk> so to login to the data centre machine as root, you switch your own desktop/laptop to text console and login there as root, sshing from root to root to get to the other machine?
[05:20] <daniels> pitti: so you have everyone with the one root account.  rad.
[05:20] <pitti> Keybuk: that's an entirely different problem; I speak about the host you login to, not the client side
[05:20] <daniels> pitti: right
[05:20] <Keybuk> pitti: no, it's exactly the same problem
[05:20] <daniels> pitti: so, at the clientside, you do it in an xterm
[05:20] <daniels> pitti: and I just send lots of XTest key events
[05:20] <daniels> and screen-scrape your xterm
[05:20] <pitti> daniels: as I said, there is no safe way to login as root under X
[05:20] <Keybuk> if you don't clean-room root the entire way, you're argument drops onto the floor and dies
[05:21] <daniels> pitti: so you are talking about the clientside
[05:21] <pitti> Keybuk: of course, if you don't trust your laptop, you shouldn't do ssh root. There is simply no other way
[05:21] <pitti> Keybuk: you HAVE to trust the whole chain
[05:21] <daniels> pitti: so what you're saying is that no-one should bother running computers?
[05:22] <pitti> Keybuk: and this is entirely orthogonal to sudo/su/whatever authentication system
[05:22] <pitti> daniels: no
[05:22] <Keybuk> pitti: no, it's entirely the same thing
[05:22] <Keybuk> as soon as there is a gateway where you go from a mortal to a superuser, that is where you attack
[05:22] <pitti> daniels: I said that the only safe way to login as root at my computer is vga text console, not from an user's session
[05:22] <Keybuk> it doesn't matter whether that's a command like su or sudo on the server, or someone's ssh client on their laptop
[05:22] <pitti> Keybuk: of course
[05:24] <Keybuk> "running anything as root is inherently insecure"
[05:24] <Keybuk> not just sudo
[05:25] <pitti> processes that run as root can't be touched by other users (apart from buffer overflows and similar holes, of course)
[05:25] <pitti> so root processes are not insecure by design, to the contrary
[05:25] <pitti> the weak point is the authentication
[05:26] <daniels> 17:22 < Keybuk> as soon as there is a gateway where you go from a mortal to a superuser, that is where you attack
[05:26] <pitti> so as soon as you use sudo and su from an user's session, root and user don't have separated contexts any more, and root is no better any more
[05:26] <Kamion> yes, and bugs that weaken the authentication further are unacceptable for exactly that reason
[05:27] <pitti> Kamion: what do you mean in particular?
[05:27] <Keybuk> pitti: but it's not just sudo and su, it's as soon as you use any method to become root from a user's session -- ssh root@ is just as attackable
[05:27] <pitti> Kamion: the sudo bug does not weaken authentication
[05:28] <pitti> Keybuk: of course, that was my exact argument
[05:28] <Keybuk> but you need to be able to become root from a user's session
[05:28] <pitti> why?
[05:28] <daniels> the sudo bug does weaken authentication
[05:28] <daniels> because it allows you to escalate privileges in ways that you shouldn't be able to
[05:28] <Keybuk> so the concern instead goes to limiting the damage
[05:28] <daniels> regardless of the default configuration, surely that's a method of weakening authentication?
[05:28] <pitti> daniels: no, it enables you to run trojan horses in an already authenticated root session
[05:29] <daniels> yes, which is escalating privileges in ways that you shouldn't be able to
[05:29] <daniels> (fwiw, xfree86 is go whenever the buildds are done)
[05:29] <pitti> daniels: if you allow users to run sudo programs without full path, you have lost anyway
[05:29] <Kamion> pitti: uh, no, restricted sudo is not a freely available root session; you're arguing that it's OK to turn restricted sudo into unrestricted sudo which it totally isn't
[05:29] <pitti> daniels: you don't need functions for that
[05:30] <Kamion> that's a rather binary argument IMHO
[05:30] <pitti> daniels: I already released xfree86
[05:30] <daniels> ah, cool.  thanks.
[05:30] <pitti> Kamion: with "restricted" you mean that you can only run some programs?
[05:30] <Kamion> yes
[05:31] <pitti> Kamion: and you say that it is safe to specify the set of allowed programs without full path?
[05:31] <daniels> (e.g. for freedesktop.org, pasc will probably be allowed to run /usr/sbin/postfix -- and *only* /usr/sbin/postfix -- as whatever he needs to, as he admins our mail)
[05:31] <pitti> daniels: ah. But pasc will _not_ be allowed to run "postfix", right?
[05:32] <pitti> daniels: if you specified /usr/bin/postfix, then you are safe even with the current sudo
[05:32] <Kamion> pitti: ah, hmm
[05:32] <azeem> does sudo qualify the path and then look it up?
[05:32] <pitti> daniels: the bug applies only if you specify without path
[05:32] <pitti> azeem: it uses $PATH normally
[05:33] <Kamion> is there a way to force a particular $PATH in sudoers
[05:33] <Kamion> ?
[05:33] <pitti> as I see it, sudo should rather clean the $PATH instead of purging all functions from the environment
[05:33] <pitti> Kamion: I have to research that
[05:34] <pitti> Kamion: you can specify the "env_reset" flag
[05:35] <pitti> Kamion: oh no, this will not alter $PATH
[05:35] <pitti> Kamion: however, it seems like a good idea anyway (and will close this export function bug, too)
[05:35] <Kamion> we build --with-secure-path="/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/X11R6/bin"
[05:35] <Kamion> so surely env_reset would reset $PATH to that?
[05:35] <pitti> ah, nice
[05:36] <pitti> Kamion: the manpage says that it only purges everything but HOME, LOGNAME, PATH, SHELL, TERM, and USER
[05:36] <pitti> Kamion: so --with-secure-path should rather be applied unconditionally
[05:36] <Kamion> yes, but then right afterwards it says that PATH will be set to the value of SECURE_PATH if sudo was compiled that way
[05:37] <pitti> right
[05:37] <pitti> currently, "sudo echo $PATH" prints for me:
[05:37] <pitti> /home/martin/bin:/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games
[05:38] <Kamion> er, $PATH is expanded *before* sudo gets a look-in in that command!
[05:38] <Kamion> <cjwatson@cairhien ~>$ echo $PATH
[05:38] <Kamion> /home/cjwatson/bin:/usr/lib/surfraw:/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games:/usr/sbin:/sbin
[05:38] <Kamion> <cjwatson@cairhien ~>$ sudo sh -c 'echo $PATH'
[05:38] <Kamion> /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/X11R6/bin
[05:38] <Kamion> that's on Debian, but should be the same
[05:38] <pitti> oh, right
[05:38] <pitti> yes
[05:39] <pitti> just checked
[05:39] <Kamion> also:
[05:39] <Kamion> <cjwatson@cairhien ~>$ which burn-cd
[05:39] <Kamion> /home/cjwatson/bin/burn-cd
[05:39] <Kamion> <cjwatson@cairhien ~>$ sudo burn-cd
[05:39] <Kamion> sudo: burn-cd: command not found
[05:40] <Kamion> so it looks like sudo already does clean the $PATH before looking up where to execute a binary, and this vulnerability is indeed relevant
[05:40] <pitti> as I said, for my sake I fix it
[05:40] <Kamion> assuming you haven't given write permissions to any of the directories in SECURE_PATH
[05:40] <pitti> well, I think this is a realatively safe assumption
[05:40] <Kamion> yep
[06:02] <azeem> lamont_r: do you guys have sbuild packages for ubuntu up somewhere?
[06:06] <lamont_r> azeem: the sbuild that we're using is almost identical to debian's, and I've been pushing diffs to neuro as I make changes.
[06:06] <lamont_r> frankly, some of the changes make no sense for debian, etc.
[06:07] <azeem> ok, so where should I hack sbuild so that it accepts warty as distribution? Or is this not possible?
[06:07] <lamont_r> oh, that diff...
[06:07] <lamont_r> sec
[06:08] <lamont_r> search for 'Bad distrib' in the file somewhere around line 576
[06:08] <lamont_r> make it look like this...
[06:08] <lamont_r>                 $main::distribution = "unstable" if $main::distribution eq "u";
[06:08] <lamont_r>                 #die "Bad distribution\n"
[06:08] <lamont_r>                 #       if !isin($main::distribution, keys(%main::dist_order));
[06:08] <azeem> heh
[06:09] <lamont_r> for extra credit, go down to after the closing brace after "Unknown option", and add:
[06:09] <lamont_r> die "Need distribution\n" if $main::distribution eq "unstable";
[06:11] <m_tthew> for extra *extra* credit, rewrite it in python.
[06:11] <lamont_r> m_tthew: actually, there's an even more invasive rewrite in progress
[06:11] <Kamion> or decent perl
[06:12] <azeem> lamont_r: other than the multibuild effort?
[06:12] <m_tthew> lamont_r: nice
[06:13] <lamont_r> sadly, the basic architectural design is completely different than debians, sadly
[06:14] <azeem> I had to touch /var/lib/sbuild/source-dependencies-warty as well, but maybe that's because I was using the packaged version of sbuild
[06:14] <infinity> lamont_r If you just pull "unstable" out of main::dist_order, then it'll happily die.
[06:14] <infinity> lamont_r : And put in the ubuntu dists too.
[06:14] <daniels> hot air on lists doesn't count as a rewrite
[06:15] <lamont_r> actually, I just changed it to require a dist, and never mind adding each new chroot I want to have...
[06:15] <lamont_r> much simpler to force me to type -dfoo
[06:15] <infinity> Perhaps.
[06:16] <lamont_r> actually, the version on my personal machine is slightly different - -dunstable still works.. :-)
[06:16] <infinity> I'd say that main::dist_order should be overridable in .sbuildrc
[06:16] <daniels> infinity: yes
[06:16] <lamont_r> given that the if statement there is the only place it's used in sbuild, I'd say it should just be deleted.....
[06:17] <azeem> will you submit a patch to debootstrap so it is possible to bootstrap Ubuntu from Debian? (in case you haven't already)
[06:18] <mirak> isn't the kernel 2.6 responsible of sound problems ?
[06:18] <lamont_r> azeem: good question for Kamion, I expect...
[06:18] <mirak> I have heard that on #debian
[06:18] <lamont_r> it's not exactly a debian bug that it lacks debootstrap scripts for other distros....
[06:19] <mirak> in fact when I rip a audio cd, I got a empty wav file
[06:19] <azeem> lamont_r: not a bug, but wishlist :)
[06:19] <infinity> lamont_r : No, but a wishlish request, certainly.
[06:19] <mirak> but the size is good
[06:19] <azeem> mirak: saying that 2.6 is broken WRT sound is a rather broad statement
[06:19] <infinity> lamont_r : If I could debootstrap warty dists from my sid machine, it'd make making chroots and playing aroiund that much simpler.
[06:19] <mirak> azeem: what does meean broad statement
[06:19] <mirak> ?
[06:20] <infinity> s/warty./ubuntu/
[06:20] <lamont_r> understood... one could fetch the warty* files from ubuntu's debootstrap, of course.
[06:20] <infinity> Oh, of course.
[06:20] <lamont_r> a wishlist would make sense
[06:20] <azeem> Log for successful build of synce-serial_0.9.0-1 (dist=warty)
[06:20] <azeem> woot
[06:20] <azeem> of course, I just figured out that synce-serial is in universe and the guy on -users probably doesn't have that
[06:21] <mirak> azeem: what means "broad statement" ?
[06:21] <azeem> mirak: it means: it is not specific enough
[06:21] <azeem> mirak: surely sound did work on Linux-2.6 and still does for most users
[06:22] <mirak> azeem: yet I can't find a reason why I can't hear audio CD sound
[06:22] <mirak> this doesn't makes sens
[06:22] <mirak> I mean that's numeric data
[06:22] <azeem> mirak: did you try #ubuntu?
[06:22] <mirak> if I can read a normal cd
[06:22] <Kamion> azeem: aj has copies of our debootstrap packages, if he wants to do so
[06:23] <mirak> azeem: I asked on every linux channel lol
[06:23] <mirak> azeem: I have heard somebody else with this problem
[06:23] <azeem> mirak: I can't help you
[06:23] <azeem> Kamion: ah, ok
[06:23] <lamont_r> mirak: the correct place for support questions is #ubuntu, this channel would be the right place to discuss your patch for fixing it...
[06:23] <mirak> azeem: I ask in dev section, because maybe, there is more "competent" people on this subject
[06:23] <azeem> mirak: the same people are also in #ubuntu
[06:24] <azeem> well, some of them at least
[06:24] <mirak> azeem: ok
[06:24] <Kamion> mirak: competent people who'll just retreat somewhere else if they don't have some peace away from support questions, though :)
[06:25] <Kamion> we used to use #ubuntu as our development channel, before the preview ...
[06:26] <mirak> Kamion: don't, I ask here, because it seems more an internal problem than a usage problem, if you see what I mean. but well I will continue to search everywhere
[06:26] <mirak> but not here :)
[06:35] <rcaskey> you know, reading debian release news these days just doesn't have the same zing as it did before ubuntu
[06:37] <azeem> what's "zing"?
[06:38] <rcaskey> excitement
[06:39] <azeem> ah
[06:55] <Kamion> anyone want to claim understanding of what's needed to fix #3770 in a udev world?
[06:58] <daniels> Kamion: fwiw, I'm pretty sure I've nailed the UTF-8 bug in libX11
[06:58] <daniels> Kamion: rebuild of that pending, right after this security build of x.org finishes ...
[06:58] <pitti> Kamion: is this really an udev problem? this rather sounds as if the alsa modules are not loaded
[06:59] <Kamion> pitti: oh, I'm not suggesting it's a udev problem
[07:03] <pitti> Can somebody please proofread https://chinstrap.warthogs.hbd.com/~pitti/usn-sudo.txt ? Thanks in advance
[07:04] <pitti> Kamion: #3770: it seems that the reporter was not using udev at all; so maybe it's just NOTWARTY?
[07:05] <Kamion> pitti: "non-fully qualified" => "non-fully-qualified", I'd say; otherwise fine
[07:05] <lamont_r> pitti: you forgot to say "for this and other reasons, one should not use a shell script that runs with elevated privileges." :-)
[07:05] <Kamion> pitti: well, it's explicitly when upgrading from sarge, which we want to support in hoary
[07:05] <Kamion> I've asked him if he installed udev and rebooted
[07:06] <pitti> Kamion: the bug report contains too little information to pinpoint the problem; but udev should solve it
[07:06] <pitti> Kamion: will upgrades from sarge to Hoary install udev?
[07:06] <Kamion> if you install ubuntu-base, yes
[07:06] <Kamion> assuming he has the alsa modules loaded ...
[07:07] <pitti> of course
[07:07] <pitti> but at least for me Sarge's hotplug should load the modules as well
[07:08] <Kamion> yes, true
[07:09] <pitti> so probably we should write "ubuntu-base" in very big letters into the upgrading instructions :-)
[07:09] <pitti> oh, gotta go... bye everybody
[07:09] <amu> lamont_r: so what we'll do ? do you setup and test the new buildsys ?  
[07:09] <lamont_r> SURE
[07:09] <lamont_r> damn caps
[07:10] <amu> cool, than just download the latest morphix-tree 
[07:11] <lamont_r> amu: could you send me an email with an rsync command? :-)
[07:11] <lamont_r> or must I mirror it the painful way?
[07:12] <amu> wget rocks ;) the other way is, ask alex for an acc   
[07:14] <Kamion> ubuntu-base> I thought we already did ...
[08:02] <RubenV> hmmm
[08:02] <RubenV> updating sed just killed my apt
[08:03] <m_tthew> RubenV : see topic
[08:03] <RubenV> aha :)
[08:03] <RubenV> thankyou
[08:03] <RubenV> ah well, this means no updates
[08:04] <RubenV> -> more time for study
[08:04] <RubenV> you should break stuff more often :)
[08:05] <usual>  cp /usr/share/info/sed.info-1.gz /usr/share/info/sed.info.gz
[08:05] <usual> fix for sed
[08:05] <usual> ?
[08:05] <usual> according to #debian
[08:05] <usual> ahh it works
[08:05] <usual> :)
[08:05] <infinity> syncing with incoming would work too.
[08:06] <usual> yea
[08:14] <Keybuk> elmo: Package base version not found: ghfaxviewer (0.22.0-4)
[08:14] <mwh> Im having some trouble getting java-applets going with mozilla-firefox, but not with mozilla and epiphany, anyone heard about this problem before=?
[08:14] <Keybuk> hmm... I really don't know what to do about that
[08:14] <Keybuk> s/^elmo/E/ heh
[08:14] <Keybuk> snapshot never archived ghfaxviewer_0.22.0-4
[08:18] <Keybuk>  E: Package base version not found: lightspeed (1.2-4)
[08:18] <Keybuk> another one too
[08:32] <kylem> sed fixed would be nice... 8)
[08:32] <Kamion> it'll happen, duded
[08:32] <Kamion> dude
[08:34] <amu> yep ;)
[08:35] <infinity> kylem : Just install the debian package from incoming on ubuntu.  Sheesh. :)
[08:36] <kylem> <- lazy
[08:36] <Kamion> wasn't an Ubuntu maintainer who broke it ...
[08:36] <kylem> oh.
[08:37] <kylem> who do i castrate?
[08:37] <kylem> :)
[08:37] <Kamion> well, Clint broke it in Debian, but it worked for him on sparc; it was a "breaks on buildds" issue
[08:37] <kylem> :\
[08:57] <mxpxpod> could someone with a cd burner test something for me on hoary?
[09:07] <ironwolf> whatcha need mxpxpod?
[09:09] <mxpxpod> ironwolf: could you download coaster's latest release, compile and run it to make sure it works on hoary?
[09:09] <mxpxpod> ironwolf: I'm not willing to upgrade to hoary if it's going to impede coaster development
[09:11] <mxpxpod> ironwolf: http://www.coaster-burn.org/files/coaster/coaster-0.1.2.tar.gz
[09:13] <ironwolf> mspxpod: what's coaster do?
[09:14] <mxpxpod> ironwolf: write cd's... data-only for now
[09:15] <ironwolf> mxpxpod: isn't that just apt-get install coaster ?
[09:15] <mxpxpod> ironwolf: I don't think it's in hoary yet....
[09:15] <lupus_> E: Couldn't find package coaster
[09:17] <mxpxpod> ironwolf and lupus_: it's not in hoary/warty... it's a project I'm working on
[09:18] <ironwolf> mxpxpod: compiling now....
[09:18] <mxpxpod> ironwolf: awesome
[09:19] <mxpxpod> ironwolf: thanks for doing this
[09:24] <fabbione> does anybody know what provides -ldml ?
[09:25] <fabbione> soryr
[09:25] <fabbione> -ldlm
[09:26] <fabbione> lamont_r: what's the problem now? ;)
[09:27] <lamont_r> xorg is d-w libglide3-dev, which fails to install (missing xorg0
[09:27] <fabbione> libglide3-dev (>= 2002.04.10-7) [alpha amd64 i386 ia64] 
[09:28] <lamont_r> on ia64
[09:28] <fabbione> but libglide3 cannot know about xorg... if there is no xorg
[09:28] <fabbione> it would pull in xfree86 build-deps
[09:28] <lamont_r>   libglide3: Depends: libx11-6 but it is not installable or
[09:28] <lamont_r>                       xlibs (> 4.1.0) but it is not going to be installed
[09:28] <lamont_r>              Depends: libxext6 but it is not installable or
[09:28] <lamont_r>                       xlibs (> 4.1.0) but it is not going to be installed
[09:28] <fabbione> E: Couldn't find package libglide3-dev
[09:29] <fabbione> this is the last error you get
[09:29] <lamont_r> of course, once xfree86 finishes building...
[09:29] <lamont_r> could just be a time thing
[09:29] <fabbione> hmmm
[09:29] <lamont_r> lib64/libdlm.so.0                                           libs/libdlm0
[09:29] <fabbione> at what phase are you now?
[09:30] <lamont_r> 3
[09:30] <lamont_r> is filling the archive
[09:30] <fabbione> make sence
[09:30] <fabbione> lvm2 is missing a build dep
[09:31] <fabbione> on libdlm0
[09:31] <fabbione> -dev
[09:31] <lamont_r> libdlm-dev
[09:31] <fabbione> yeah that one
[09:32] <fabbione> ok.. bug has been reported already
[09:32] <fabbione> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=281433
[09:32] <fabbione> this is interesting :-)
[09:36] <fabbione>  ./main/binary-sparc/: New 1314kB 1129 files 496MB 11s
[09:36] <fabbione>  ./main/source/: 2 pkgs in 0s
[09:36] <fabbione> i love this :P
[09:37] <fabbione> lamont_r: unknown/silo_1.4.8-1 [-:uncompiled] 
[09:38] <fabbione> what was the trick to switch from unknown to whatever it really is?
[09:39] <lamont_r> fabbione: overrides file, or fix the ^*_& package
[09:39] <fabbione> lamont_r: ok :-)
[09:39] <spotter> seb128: the evolution panel problem isn't fixed
[09:39] <fabbione> thanks
[09:40] <spotter> it now just references evolution-2.2
[09:40] <lamont_r> fabbione: it
[09:40] <spotter> it should be just evolution
[09:40] <lamont_r> s an archive thing, not a buildd thing
[09:40] <spotter> what happens w/ evolution-2.4
[09:40] <seb128> spotter: there is no evolution.desktop
[09:41] <fabbione> lamont_r: yes.. i just didn't bother to create an override file
[09:41] <lamont_r> me neither. :-)
[09:41] <fabbione> lamont_r: it was just to schedule it before other packages
[09:41] <spotter> seb128: there should be, and that should be what's in the panel by default
[09:41] <spotter> otherwise upgrades are going to be a bitch w/ each release
[09:42] <lamont_r> could always shut down the buildd, take it, and then manually sbuild it before starting back up the buildd... :-p
[09:42] <seb128> spotter: not sure
[09:42] <fabbione> lamont_r: ehehe
[09:42] <fabbione> lamont_r: that's cheating :P
[09:42] <spotter> just my not so humble opinion, take it as you will.
[09:42] <lamont_r> fabbione: nah - that's _EFFICIENCY_ :-)
[09:42] <fabbione> rotlf
[09:42] <spotter> seb128: I'm just annoyed w/ it as its constantly crashing on me
[09:43] <fabbione> time to get some food
[09:43] <fabbione> thanks lamont
[09:43] <fabbione> later fellas
[09:43] <seb128> spotter: the problem is the way to handle the // installation of differents version of evolution
[09:43] <lamont_r> note also that our wanna-build isn't necessarily getting universe/multiverse (if present) after main, even.  - the DC wanna-build has the config to do that
[09:43] <seb128> spotter: like 1.4 and 1.5 were // installable, so have only one .desktop is not optimal
[09:44] <spotter> seb128: the way I view it
[09:44] <spotter> evolution.desktop == default
[09:44] <spotter> i.e. install 2
[09:44] <spotter> evolution-2.2, which is in menu and autobuilt
[09:44] <spotter> hmm
[09:44] <spotter> never mind
[09:45] <spotter> debian seems to handle it fine
[09:45] <seb128> debian doesn't have a launcher in the panel :p
[09:45] <spotter> default evo (2.0) installs as evolution
[09:45] <seb128> same here 
[09:45] <spotter> seb128: so defualt evolution.desktop should be what ever is your default
[09:45] <spotter> if you want another evo installed, it installs a versioned evo
[09:45] <spotter> .desktop
[09:46] <spotter> i.e. on my laptop b4 it died I had 2.0 and 2.2 runnable
[09:46] <seb128> you want to set an alternative on the desktop file ?
[09:46] <spotter> no alternative
[09:46] <seb128> there is no other way to have the same file in 2 packages
[09:46] <spotter> why do you need the same file in 2 packages?
[09:46] <spotter> are there 2 evo's installable in ubuntu?
[09:46] <seb128> not atm, but that could happen
[09:47] <spotter> ok, so you just control which is considered default
[09:47] <spotter> if user wants to use non default, that's their choice
[09:47] <seb128> how do you set the default ?
[09:47] <spotter> you do
[09:47] <spotter> aka ubuntu
[09:47] <seb128> ie: 1.4 and 1.5 .. 1.4 has the evolution.desktop
[09:47] <spotter> yes, as that was stable
[09:47] <seb128> and then 1.5 -> 2.0 and 2.0 should have the .desktop
[09:47] <spotter> when 1.5 went stable to 2.0, 1.4 disappeared
[09:48] <spotter> i.e. there's evolution and evolution-unstable (aka evolution2.2)
[09:48] <seb128> yeah, could be in this way
[09:48] <fabbione> lamont_r: yes i know. i configured only main atm
[09:48] <seb128> much easy to just change the panel link :)
[09:48] <fabbione> (food isn't ready yet)
[09:48] <lamont_r> fabbione: yeah
[09:48] <lamont_r> ia64 buildd's are into universe
[09:48] <seb128> spotter: what's the big deal to have the default in the panel ?
[09:48] <spotter> seb128: but sucks for users ugrading who dont know and just know after an upgrade it doesnt work anymore
[09:49] <fabbione> lamont_r: as i was writing yesterday at the meeting, i don't have enough processing power for universe
[09:49] <spotter> seb128: when users upgrade from warty to hoary, their evolution panel icon wont work anymore
[09:49] <seb128> spotter: yeah, but hoary is a devel branch, don't worry it'll be fine before the release whatever is the fix
[09:49] <spotter> unsure it's possible to fix w/o ugliness
[09:50] <spotter> what I'm proposing is a way to think about this for the future
[09:50] <spotter> no .desktop entry should have a versioned filename
[09:50] <seb128> some people want 2.1 in the panel/menu
[09:51] <seb128> why not having both version here if both version are installable in // ?
[09:51] <spotter> users can still do that
[09:51] <fabbione> lamont_r: but if the port will be accepted, Mark will buy buildds
[09:51] <fabbione> ok now food is ready
[09:51] <spotter> but ubuntu's supported is just one
[09:51] <fabbione> bye :-)
[09:51] <spotter> i.e. does ubuntu support courier?
[09:51] <spotter> even though that can be installed
[09:51] <lamont_r> fabbione: right
[09:51] <spotter> they support postfix
[09:55] <jdub> morning
[09:57] <Keybuk> jdub: Good Morning.  Today is Boomtime, the 30th day of The Aftermath in the YOLD 3170.  You have no new messages.  You have no appointments.
[09:58] <lamont_r> Today is Sweetmorn, the 29th day of The Aftermath in the YOLD 3170
[09:58] <lamont_r> Keybuk: wassup with your calendar?
[09:58] <Keybuk> lamont_r: it's the 17th today ...
[09:58] <lamont_r> same here
[09:58] <Keybuk> actually, for jdub it's the 18th
[09:58] <lamont_r> doh
[09:59] <Keybuk> :)
[10:00] <mako> people are totally unjustified in calling ubuntu the "debian-based distro with the funny name"
[10:01] <mako> http://distrowatch.com/dwres.php?resource=independence
[10:02] <mako> BEERnix?
[10:02] <mako> overclockix?!
[10:03] <Keybuk> "Debian" itself is a pretty funny name, when you think about it
[10:03] <Keybuk> people are just used to it
[10:03] <mako> Keybuk: not nearly as funny as beernix :)
[10:03] <mako> or "penguin sleuth"
[10:03] <Keybuk> In fact, Linux is reasonably silly
[10:04] <Keybuk> (and has equal problems with people pronouncing it correctly :p)
[10:04] <mako> right, i think ubuntu belongs in the class of debian and linux
[10:04] <mako> basically whimsical sounding names that gain a meaning and seem natural later for the vast majority of people
[10:04] <Keybuk> at least "Ubuntu" is actually a word in a widely-spoken language
[10:04] <mako> i have now seen ubuntu spelled with every possible vowel in the last letter
[10:04] <jdub> "widely"
[10:04] <mako> jdub: compared to cornish, sure
[10:04] <Keybuk> of course, they could be referring to "Warty Warthog" as the funny name
[10:05] <Keybuk> and they've got us by both balls, there
[10:05] <mako> linux loco!
[10:05] <mako> or, BIG LINUX
[10:05] <mako> (all caps)
[10:06] <Keybuk> FAT LINUX!
[10:06] <mako> yeah, they're just inviting a BIGGER LINUX with that name
[10:06] <spotter> mako: did you ever find a place to host that event you asked the acm about?
[10:06] <mako> spotter: nope, it was too bad
[10:06] <mako> spotter: it was super last minute
[10:06] <mako> spotter: it was a bad combination of me just moving here and it being last minute
[10:06] <spotter> so it wasn't held?
[10:07] <mako> spotter: no organized talk
[10:07] <spotter> so you live at the med school?
[10:07] <mako> spotter: yep
[10:07] <spotter> got to love the heights
[10:07] <mako> spotter: you are at columbia?
[10:07] <spotter> are
[10:07] <spotter> argh
[10:08] <shaya> that would work now
[10:08] <mako> true enough, morningside?
[10:08] <shaya> that's where school is, home is washington
[10:08] <mako> you coming tonight?
[10:08] <mako> shaya: dc or state?
[10:08] <shaya> heights
[10:08] <mako> haha
[10:08] <mako> ok, i'm from washington STATE, living in washington HEIGHTS
[10:09] <mako> i do love washington heights
[10:09] <mako> you coming to the talk tonight?
[10:09] <shaya> i'm from dc suburbs, spent summer in the state working for the man
[10:10] <mako> shaya: www.nylug.org
[10:10] <shaya> yes, but you go to restaurants
[10:10] <shaya> problematic for me
[10:11] <mako> shaya: it may be too late to rsvp for their (slightly authoritarian) rsvp system
[10:11] <shaya> yes
[10:11] <shaya> it is
[10:11] <shaya> saw your blog post
[10:11] <shaya> it was closed when you posted it
[10:11] <mako> blast
[10:11] <mako> there is this "stammtisch" thing
[10:11] <shaya> stammtisch?
[10:11] <shaya> sounds like a yiddish word
[10:11] <mako> at 8:30.. keysigning and hanging out in the bar across the street
[10:12] <mako> last monght i missed the meeting but just went to the hang-out
[10:12] <mako> last month even
[10:12] <shaya> http://yucs.org/~spotter/scary-me.jpg
[10:12] <shaya> come dressed like that and I wouldn't fit in :)
[10:13] <mako> i think it would be fine
[10:14] <mako> but if you don't make it tonight, we can get together some other time if you'd like
[10:14] <mako> there is a talk at the morningside campus i'm going to tomorrow on free software stuff
[10:14] <mako> http://www.columbia.edu/cu/arts/dmc/docs/lectureseries.html
[10:30] <seb128> jdub: here ?
[10:34] <Cube-ness> any estimation on when sed will be fixed?
[10:36] <seb128> soon
[10:38] <Cube-ness> hehe.. cool
[10:40] <sivang> Cube-ness :  Quick workaround for the sed problem in hoary: edit /var/lib/dpkg/info/sed.postinst as root, add "exit 0" in the second line...
[10:42] <ironwolf> or add warty to sources.list and apt-get install sed/warty ;)
[10:45] <Cube-ness> ok
[10:46] <hornbeck> sivang: Under the #! /bin/sh -e?
[10:46] <hornbeck> or under pkg=sed
[10:46] <sivang> under pkg=sed,
[10:46] <sivang> worked for me
[10:46] <hornbeck> ok thanks
[10:47] <hornbeck> that would be the fourth line
[10:47] <hornbeck> sivang: thanks that just helped me :-)
[10:48] <sivang> hornbeck : no prob.
[10:52] <seb128> elmo_away: gpdf sync please
[10:55] <Keybuk>      ^^^^
[10:56] <Keybuk> IRC is a lossy protocol, you should probably IM him or send him a mail :p
[10:56] <lamont_r> Keybuk: but elmo_away is all seeing. :-)
[10:57] <lamont_r> although elmo_away may be his home instance, and that'll be really slow to respond... :(
[10:58] <lamont_r> prolly
[11:00] <Keybuk> I fear anyone who can keep up with scrollback these days
[11:00] <Keybuk> I know elmo's bitched in the past about people leaving messages on channel that he's never seen
[11:00] <lamont_r> yeah
[11:01] <lamont_r> it's good for messages that you don't really care if they get, but not good for critical messages (those I send a follow-up email for..)
[11:04] <Cube-ness> question about sound in hoary.. while my laptop works fine, this machine makes me kill esd in order for oss sound to work.. is there a better fix?
[11:05] <Keybuk> Cube-ness: 1) ask in #ubuntu   2) use esd instead of oss sound
[11:06] <Cube-ness> well, a lot of apps "just use" oss
[11:06] <Cube-ness> i asked here because i assumed folks in devel would be more familiar with hoary at this point in time
[11:07] <lamont_r> Cube-ness: #ubuntu has _lots_ of devel folks
[11:07] <Cube-ness> dont worry, i am not gonn aflood his place with general usage questions.. hehe
[11:07] <lamont_r> -devel is for discussing code changes
[11:07] <Cube-ness> ok
[11:07] <Keybuk> Cube-ness: developers hang out on #ubuntu too ... this channel is for discussion about development, not a place to find the developers
[11:07] <Cube-ness> right
[11:08] <Keybuk> and the truth is even the developers aren't all-knowing, sometimes a fellow user will have a much better suggestion
[11:14] <amu> lamont_r: #3046 should i close it ?   
[11:15] <Keybuk> lamont_r: how far along is it? :p
[11:16] <lamont_r> Keybuk: waiting for sync from debian
[11:17] <lamont_r> amu: I expect
[11:17] <lamont_r> with comments, of course.
[11:17] <jdub> seb128: hey hey hey
[11:17] <Keybuk> ftp.debian.org still only has 4.1.2-4
[11:17] <lamont_r> like "mirroring the latest packages from morphix and running the script in the chroot as non-root, but allowed sbuild" type commnets
[11:17] <lamont_r> Keybuk: the buggy version is 4.1.2-2
[11:18] <Keybuk> so when did -3 or -4 pop into unstable?
[11:18] <Keybuk> tonight's dinstall run?
[11:18] <amu> lamont_r: thats another one :) he asked about QTparted on the liveCD :)  
[11:19] <lamont_r> ??
[11:19] <lamont_r> yes
[11:19] <lamont_r> Keybuk: yes
[11:19] <amu> just look to #3046  
[11:19] <Keybuk> elmo's sync runs "daily" I think ... so we should've seen that by now, no?
[11:20] <lamont_r> it stalls until after the debian mirrors have time to finish syncing the pulse.
[11:20] <seb128> jdub: why not adding a conflict between gamin and fam ?
[11:20] <lamont_r> hence "any time now"...
[11:20] <lamont_r> not sure when the sync script runs
[11:20] <Keybuk> hmm, last run I see was 22:12
[11:20] <lamont_r> utc?
[11:20] <Keybuk> Jackass HTTP Time :p
[11:20] <lamont_r> jht.  got it.
[11:20] <lamont_r> so 8 min ago
[11:21] <Keybuk> Last-Modified: Wed, 17 Nov 2004 22:12:20 GMT
[11:21] <Keybuk> (on the file it spits for mom)
[11:22] <jdub> seb128: because the packaging death squad would have my head :)
[11:23] <Keybuk> jdub: can you install them both at the same time?
[11:23] <seb128> jdub: what's the problem with the conflict ? For the moment we only get users with broken system because they have both installed
[11:23] <jdub> Keybuk: yes
[11:23] <jdub> seb128: no, they only get a broken system if they have fam+libgamin
[11:23] <Keybuk> if you install them both at the same time, can apps use one and other apps use the other?
[11:23] <jdub> Keybuk: no
[11:23] <jdub> Keybuk: libgamin + fam == broken
[11:24] <Keybuk> why is it broken?
[11:24] <jdub> because they don't talk to each other correctly
[11:24] <Keybuk> does libgamin not depend on gamin?
[11:24] <seb128> jdub: fam+gamin with gamin->libgamin0 and conflicts on libfam... you get fam+gamin+libgamin
[11:24] <jdub> libgamin recommends gamin
[11:24] <seb128> yeah, but they install gamin, get libgamin0 but fam is not removed
[11:24] <Keybuk> if you have libgamin, gamin and fam installed -- is it broken?
[11:24] <seb128> and BOUM
[11:24] <seb128> yes
[11:24] <jdub> seb128: fam+gamin+libgamin ought to work...?
[11:25] <seb128> jdub: dunno if it "ought" but it doesn't
[11:25] <jdub> Keybuk: if you have libgamin and fam, that is broken to start with
[11:25] <seb128> according to #ubuntu
[11:25] <jdub> seb128: ok
[11:25] <Keybuk> jdub: but they won't get gamin installed anyway, because the dependency isn't strong enough *shrug*
[11:25] <jdub> Keybuk: yes.
[11:25] <seb128> but what's the problem with the conflict between gamin and fam ?
[11:26] <Keybuk> seb128: they don't conflict.
[11:26] <jdub> seb128: because by definition of the rules, they don't actually 'conflict'
[11:26] <Keybuk> they just happen to break each other.
[11:26] <shaya> jdub: try installing kde
[11:26] <seb128> Keybuk: gamin depends on libgamin0 and gamin+libgamin0+fam=broken ... seems to be a conflict case for me
[11:26] <shaya> it's in universe, but a pita to install from aptitude
[11:26] <Keybuk> seb128: no, that's not a conflict.
[11:26] <jdub> shaya: why?
[11:26] <shaya> it complains for some reason
[11:26] <Keybuk> conflict is "you must utterly remove this package before you even unpack me"
[11:27] <seb128> Keybuk: I've no problem to use a conflit to avoid breakages :p
[11:27] <shaya> try installing konq frm aptitude and you'll see
[11:27] <jdub> shaya: why should i try installing it? :)
[11:27] <shaya> dont install it
[11:27] <shaya> just try in aptitude
[11:27] <jdub> shaya: i don't use aptitude, we don't support kde...
[11:27] <Keybuk> seb128: and you're one of the primary reasons that apt and dpkg sometimes hang or fuck up doing upgrades :p
[11:27] <lamont_r> keybuk: if it synced at :12, then it should show in w-b in about 6 min, and be in the archive in 36
[11:27] <shaya> jdub: true, but you do provide it in universe
[11:27] <seb128> jdub: they'll get me first apparently :p
[11:27] <shaya> one would think stuff in universe should be installable
[11:28] <jdub> shaya: universe == unsupported -> if there are bugs, they're not on my priority list
[11:28] <shaya> that's a different story
[11:28] <seb128> Keybuk: right ... but what's the right solution in this case ? I'm not forcing an upgrade this time :)
[11:28] <Keybuk> what's in universe, built.  there's no other guarantees
[11:28] <shaya> i had no interest in kde, just ran into it as needed kde-dev stuff to build redhat-artwork
[11:28] <lamont_r> shaya: if there's something in universe that is not built or not installable, we welcome patches....
[11:29] <shaya> what can I say, I like bluecurve
[11:29] <lamont_r> shaya: likewise, if it's circular build-deps holding it up, please let me know...
[11:29] <jdub> seb128: i'll double-check with fam/gamin/libgamin
[11:29] <Keybuk> seb128: someone actually sitting down and working out what's wrong, would be a good start
[11:29] <lamont_r> kde is known to be b0rked on ppc, but is there on i386/amd64
[11:29] <shaya> lamont: I was really confused.  aptitude refused to install it, but a plain apt-get install worked fine (without removing anything)
[11:29] <lamont_r> interesting - could be a recommended package that's not installable, then
[11:29] <Keybuk> shaya: aptitude does that sometimes ... it's a way
[11:29] <seb128> Keybuk,jdub: is there any interest to get gamin AND fam installed ?
[11:30] <lamont_r> aptitude (IIRC) likes to install recommended packages by default
[11:30] <lamont_r> apt-get doesn't.
[11:30] <Keybuk> which is seeded?  fam or gamin?
[11:30] <shaya> here's something else weird
[11:30] <shaya> for konqueror
[11:30] <shaya> Depnds: ....  kfind
[11:30] <shaya> but
[11:30] <jdub> seb128: no
[11:30] <jdub> Keybuk: gamin
[11:30] <shaya> hmm
[11:31] <Keybuk> jdub: are either libgamin or libfam seeded?
[11:31] <shaya> not anymore
[11:31] <shaya> weird
[11:31] <Keybuk> does gamin depend on libgamin?
[11:31] <jdub> Keybuk: don't think so
[11:31] <jdub> yes
[11:31] <Keybuk> ok
[11:31] <jdub> if you have ubuntu-desktop, you'll get gamin
[11:31] <Keybuk> so anyone with ubuntu-desktop has gamin and libgamin
[11:32] <jdub> but if you upgrade without ubuntu-desktop, you only get libgamin
[11:32] <jdub> i know that libgamin+fam is utterly b0rk
[11:32] <Keybuk> libgamin functionally-replaces libfam, yes ?
[11:32] <jdub> i'm just about to test libgamin+gamin+fam
[11:32] <jdub> yes
[11:32] <jdub> it's a binary compatible subset
[11:32] <Keybuk> does gamin functionally-replace fam?
[11:32] <jdub> roughly, but not really :)
[11:32] <seb128> jdub: the trashapplet (at least) is broken with this according to some user
[11:33] <Keybuk> do applications depend on libgamin?
[11:33] <jdub> gam_server is a user daemon
[11:33] <Keybuk> or do they depend on "something that provides libfam"
[11:33] <jdub> fam is a global daemon, which can communicate to the outside world, too
[11:33] <jdub> atm, gnome-vfs depends on either
[11:34] <Keybuk> is libgamin useful without gamin?
[11:35] <jdub> to let your program link, sure :)
[11:35] <Keybuk> that's not what I asked
[11:35] <jdub> but it survives without gamin
[11:35] <Keybuk> is there any reason to have libgamin installed without gamin?
[11:35] <jdub> generally no
[11:35] <jdub> but it depends if you're smoking debian crack or not
[11:35] <Keybuk> then why doesn't it depend on gamin?  was somebody getting wishy-washy about dependencies? :p
[11:36] <jdub> because it doesn't strictly depend on gamin
[11:36] <Keybuk> "it absolutely critically needs this package, but you *might not* use that functionality, so it's just a Recommends"
[11:36] <jdub> same way that libfam doesn't strictly depend on fam
[11:36] <jdub> i just explained that it doesn't absolutely critically need the package
[11:36] <Keybuk> libgamin works without gamin?
[11:36] <Keybuk> you can use its functionality without gamin installed?
[11:36] <jdub> 09:35 < jdub> to let your program link, sure :)
[11:36] <jdub> 09:35 < jdub> but it survives without gamin
[11:37] <jdub> exactly the same as the libfam/fam relationship
[11:37] <Keybuk> so an application can use libgamin without gamin installed?
[11:37] <jdub> it will link, yes
[11:37] <Keybuk> I didn't ask whether it'd link :op
[11:38] <jdub> that's all it does, dude
[11:38] <jdub> "using" libgamin is linking to it
[11:38] <Keybuk> so what does libgamin do?!
[11:38] <jdub> libgamin talks to gamin
[11:38] <jdub> that's what it does
[11:38] <jdub> software links to libgamin
[11:38] <seb128> :)
[11:38] <Keybuk> sure, you call functions in libgamin to use gamin?
[11:38] <jdub> yes
[11:38] <Keybuk> so if you use libgamin, you *need* gamin installed?
[11:38] <jdub> 09:36 < jdub> 09:35 < jdub> but it survives without gamin
[11:38] <jdub> 09:37 < jdub> exactly the same as the libfam/fam relationship
[11:38] <Keybuk> libgamin should depend on gamin
[11:39] <Keybuk> libfam should depend on fam
[11:39] <Keybuk> libfam and libgamin functionally-replace each other
[11:39] <Keybuk> problem solved
[11:39] <lamont_r> jdub: so when do we get a good gnome audio burner?? huh? huh? huh???
[11:39] <Keybuk> if you weren't being so wishy-washy-liberal about your dependencies
[11:39] <jdub> i've done what was done with fam
[11:39] <ironwolf> lamont_r: I'm working on it.
[11:39] <Keybuk> "you don't *need* to eat your dinner ... you could just look at it, sure you'll die, but you don't *need* to eat"
[11:39] <jdub> because, in debian mode, this is not a dependency
[11:39] <Keybuk> it so is
[11:39] <jdub> dude
[11:39] <Keybuk> I'd make those dependencies in Debian
[11:40] <jdub> YOU DO NOT NEED GAMIN
[11:40] <Keybuk> there are far too many Recommends, again especially in the GNOME and KDE camps, which should be Depends.
[11:40] <jdub> lamont_r: coaster's doing something about it, which we may get in hoary
[11:40] <Keybuk> ok, so if you don't need gamin, it doesn't matter that it breaks :p
[11:41] <jdub> that's not what breaks
[11:41] <jdub> if you'd been listening
[11:41] <seb128> jdub: I don't get the problem with nautilus beeing slow without gamin ... which part is slowing it ?
[11:41] <Keybuk>   The `Depends' field should be used if the depended-on package is
[11:41] <Keybuk>   required for the depending package to provide a significant
[11:41] <Keybuk>   amount of functionality.
[11:41] <jdub> seb128: i'll figure out what's going on with libgamin/fam/gamin, and get back to you
[11:41] <seb128> jdub: ie: is it a nautilus or gamin issue ?
[11:42] <seb128> jdub: ok, thanks
[11:42] <jdub> seb128: libgamin issue, i believe
[11:42] <Keybuk> I'm not trying to be a pain here, I'm actually trying to help
[11:42] <seb128> ok, I think so, nautilus has not changed and it used to work fine with fam
[11:42] <Keybuk> conflicting something as low-down in the chain as fam could result in APT/dpkg needing to deinstall half the packages just to install gamin
[11:42] <seb128> speaking about nautilus I've tested the bonobo-slay branch today, it's almost fully working :)
[11:43] <jdub> seb128: rocking :-)
[11:43] <jdub> seb128: was going to ask ;)
[11:43] <jdub> it's going to be sweet
[11:43] <Keybuk> conflicts are not fluffy, they are brutal
[11:44] <Keybuk> gamin could Conflict&Replace&Provide fam
[11:44] <Keybuk> or you could up the Recommends to Dependencies, so force fam's removal that way
[11:45] <seb128> how that forces the removal ?
[11:45] <Keybuk> which?
 or you could up the Recommends to Dependencies, so force fam's removal that way
[11:46] <Keybuk> gamin is seeded, so would be installed as a dependency of ubuntu-desktop
[11:46] <seb128> you still need a Conflicts right ?
[11:46] <Keybuk> libgamin0 would be installed as a dependency of gamin
[11:46] <Keybuk> libfam0 would be removed as a Conflict&Replace&Provide of libgamin0
[11:46] <Keybuk> fam would be removed as it depends on libfam0
[11:46] <seb128> ok, you need the Conflict&Replace&Provide of libgamin0
[11:46] <sivang> seb128 : could you please apply the fix to yelp (#3557) so doc team people could use yelp on hoary? or do you prefer we'd just fetch the package from warty?
[11:46] <Keybuk> no Conflict.
[11:47] <Keybuk> (Conflict&Replace&Provide is magic, it's not an ordinary Conflict)
[11:47] <seb128> Keybuk: better to Replaces&Provide too than only Conflict ?
[11:47] <seb128> oh ok
[11:47] <Keybuk> seb128: hell yes.
[11:47] <seb128> I'll remember that for the next time :)
[11:47] <Keybuk> Conflicts & Replaces means "absolutely replaces the functionality of"
[11:48] <Keybuk> Conflicts: foo
[11:49] <Keybuk> means that you must prerm, remove and postrm 'foo' (utterly remove) before you even *unpack* the package
[11:49] <Keybuk> given that APT/dpkg generally unpack everything at once, this is bad
[11:49] <Keybuk> whereas
[11:49] <Keybuk> Conflicts: foo
[11:49] <Keybuk> Replaces: foo
[11:49] <Keybuk> Provides: foo
[11:49] <Keybuk> is a "total replacement for foo", it can be safely swapped at any point
[11:50] <sivang> seb128 : (I remember you told me that you might prefer to wait for 2.10 :)
[11:51] <seb128> hum ok, good to know
[11:51] <seb128> Keybuk: thanks :)
[11:51] <seb128> sivang: I was hoppy for a 2.9.2 upstream soon in fact, but if somebody need it I'll apply the patch tomorrow
[11:52] <Keybuk> but that's a per-package thing... Provides: foo (<< 1.2) is illegal :p
[11:52] <sivang> seb128 : thank you!
[11:52] <seb128> usually I only use Replaces/Provides for name changes
[11:52] <seb128> and brutal Conflict for something like fam/gamin
[11:53] <seb128> I'll try to think about this next time :)
[11:53] <Keybuk> Conflict on its own means one thing, and one thing only.  The two packages contain a file of the same name.
[11:53] <Keybuk> That's it.
[11:53] <Keybuk> Using it for *anything* else, will produce bad behaviour.
[11:54] <seb128> I used to use that as a "can't work together"
[11:54] <jdub> i did a c/r/p for polypaudio
[11:54] <jdub> because it has an esd file in it :)
[11:54] <jdub> and really does functionally c/r/p esound
[11:54] <Keybuk> jdub: yeah, I'd be happy if gamin c/r/p fam
[11:55] <Keybuk> seb128: that's a *really* bad thing to do, because conflicts act before unpacking -- they're as evil as pre-depends
[11:55] <Keybuk> I'm going to RSN add "Breaks" to dpkg -- which is a field for "doesn't play well together"
[11:58] <seb128> yeah, I don't really know how apt internals work, thanks for the explanations
[11:58] <jdub> so the consensus is c/r/p fam in gamin?
[11:59] <seb128> somebody should write a "why Conflicts is bad" document, I'm sure that a lot of the maintainers don't know a lot about how apt handle that :)