[00:00] <slangasek> no, hppa just doesn't have someone willing to give it a dignified end
[00:00] <slangasek> it's been in worse shape than alpha for longer
[00:00] <NCommander> Using linuxthreads on glibc 2.7 :-P
[00:01] <NCommander> (which is what the m68k plan is if our compiler ever gets to the state that it has __thread)
[00:06]  * NCommander kicks ubuntu-cd alive
[00:09] <superm1> slangasek, so as to make sure that those changes to the seeds don't forget forgotten prior to alpha5, is there a particular place you'd like to see seed bugs filed at?
[00:10] <slangasek> superm1: hum, there ought to be a place, but I don't know what it is
[00:10] <slangasek> cjwatson: superm1 pointed out that nvidia drivers aren't included on the DVD currently; is there any reason you know of that they shouldn't be?
[00:16] <cjwatson> slangasek: not that I know of; maybe seed tweaks needed after they got split out from linux-restricted-modules
[00:16] <cjwatson> superm1: historically we've stuck seed bugs on ubuntu-meta
[00:30] <SJobs> is there someone who is great with kernel so they can check it out and reverse engineer how the hyper-visor(ps3) communicates with it and see if you can find exploits ?
[00:33] <NCommander> sbeattie, ok, GNOME is installed, and dinner was eaten, testing now
[00:33] <superm1> cjwatson, okay thanks.  i've filed a bug  there
[00:35] <calc> is it normal for lintian on ubuntu to not know about Maintainer field differing from Changelog?
[00:35] <calc> i notice it is warning a lot about that while repackaging OOo
[00:35] <NCommander> sbeattie, confirming original bug works as expected
[00:36] <calc> ah maybe its because i versioned the package incorrectly, heh
[00:36]  * calc adds ubuntu1 to the package name
[00:37] <NCommander> calc, why are you repackaging OOo?
[00:38] <calc> NCommander: split source packages
[00:38] <NCommander> ah
[00:38] <wgrant> So it won't take a decade to build?
[00:38] <calc> yep ubuntu1 fixed it
[00:38] <NCommander> YAY
[00:39] <NCommander> It took three weeks to build on m68k
[00:39] <calc> wgrant: it will still take a decade to build but there will be like 20 source packages instead
[00:39] <NCommander> I think we FINALLY added that beast to NFU
[00:40] <wgrant> calc: But it will no longer be like KDE where a typo fix in some documentation makes me download 60MB of icons again?
[00:40] <calc> wgrant: probably not, at least that is the hope :)
[00:40]  * NCommander hugs calc
[00:40] <calc> of course OOo is still in early stages of being split so this may not make it for intrepid, not sure yet
[00:41] <calc> its being split in go-oo not upstream, at least not yet
[00:41] <NCommander> BenC, ping
[00:41] <NCommander> calc, if you need a beta tester, I'll help
[00:42] <calc> NCommander: when i get it far enough along to be usable for testing i'll post on the relevant lists/forums
[00:42]  * calc is still working on getting the first source to build properly
[00:42] <NCommander> calc, PPAs are your friends
[00:42] <NCommander> If you run into FTBFS issues, hit me up
[00:42] <NCommander> I'm fairly good at resolving them
[00:42] <calc> NCommander: yep i have 3.0b2 in there now
[00:43] <calc> the main issue is that i am converting a 3600 line rules file and getting it to build at the same time :)
[00:43] <NCommander> Ow
[00:43] <calc> NCommander: oh btw i used to maintain kde for debian (long ago)
[00:43] <NCommander> 3600 LINE?!
[00:43] <calc> well 3599 lines actually :)
[00:43] <NCommander> *blinks*
[00:43] <NCommander> For the love of god ...
[00:43] <NCommander> I've built OOo from source
[00:43] <calc> so far the new rules is only 409 lines
[00:43] <NCommander> How the heck can that beast be 3600?
[00:44]  * NCommander discovers the Ubuntu-smokers group .
[00:44] <calc> NCommander: OOo is very complicated :)
[00:44] <NCommander> I realize that
[00:44] <NCommander> But 3600?
[00:44] <NCommander> Christ, there are packages in Debin that are smaller that including all the source in the orig.tar.gz
[00:44] <NCommander> (not many, but sitll)
[00:44] <calc> 409 lines for the new split package is just very basic support for the configure options, not much of anything in it yet
[00:45] <NCommander> sbeattie, can't confirm the fix
[00:45] <NCommander> Doesn't work right
[00:45] <sbeattie> bugger.
[00:46] <NCommander> sbeattie, its now recongizing archive.ubuntu.com as a third party software
[00:46] <calc> http://pastebin.ubuntu.com/40799/
[00:46] <calc> stuff like that
[00:46] <NCommander> Sources still pull from there, so it needs to recongize both
[00:46] <sbeattie> gar.
[00:46] <NCommander> let me post my sources.list
[00:47] <NCommander> sbeattie, I lost the bug, please relink
[00:47] <calc> i'm trying to clean up as i go along since the current rules has evolved over 6 years and not really cleanly in some ways
[00:47] <sbeattie> Launchpad #220890
[00:48] <sbeattie> NCommander: is what you're seeing the same as bug 244093 or something different?
[00:50] <NCommander> You mean security.ubuntu-ports.org in third-party software?
[00:50] <sbeattie> yeah, something like that.
[00:51] <NCommander> Nope
[00:51] <NCommander> Just archive.ubuntu.com is showing up in third party
[00:51] <NCommander> actually
[00:51] <NCommander> Also the CDs are
[00:51] <calc> all that use system stuff in OOo should be default, they ship copies of all those libs in the monolithic source (gag)
[00:53] <sbeattie> NCommander: I have to run. Can you report your results as a comment to bug 220890? Thanks for testing it out!
[00:53] <NCommander> sbeattie, I also uploaded a screenshot
[00:53] <sbeattie> excellent, thanks!
[00:54] <NCommander> I'm subscribed, so if any more testing needs to be done, hit me up
[01:10] <StevenK> cjwatson: It's a copy of the older dailies. I didn't want to delete them in case someone missed them. It looks like no one has ...
[01:57] <lifeless> EFAILTOUNDERSTANDUPGRADES: https://bugs.launchpad.net/bugs/75073
[03:49] <poopcheese> I ASK THAT THE UBUNTU DEVELOPERS REMOVE ALL PROGRAMS THAT USE MONO.
[03:50] <StevenK> Joining an IRC channel and yelling it isn't way to get it done.
[03:50] <wgrant> I ask that Mono trolls don't shout nor troll.
[03:50] <poopcheese> for me it is
[03:50] <poopcheese> microsoft could sue the devels
[03:51] <poopcheese> REMOVE IT
[03:52]  * Hobbsee requests that poopcheese does something constructive.
[03:52] <poopcheese> PLUS IT IS MADE MY NOVELL
[03:52] <poopcheese> YOUR MEAN
[03:52] <wgrant> I didn't know she had one.
[03:52] <poopcheese> =/
[03:52] <ssweeny> the caps lock key is not an Awesome Button
[03:53] <ion_> wgrant: I borrowed my mean to her.
[03:53] <Hobbsee> poopcheese: mono will not be removed at this time.
[03:53] <poopcheese> WHY NOT
[03:53] <RAOF> ssweeny: No, but it is CRUISE CONTROL FOR COOL!
[03:53] <poopcheese> IT IS A PIG
[03:53] <wgrant> RAOF: This is true.
[03:53] <Hobbsee> poopcheese: and drop the caps.
[03:54] <poopcheese> SRY MY KEYBOARD IS LIKE THIS
[03:54] <ssweeny> poopcheese, you're free to apt-get remove mono and any programs that use it
[03:55] <poopcheese> HAHA I DONT HAVE MONO ON MY COMPUTER IN FACT I USE WINDOWS XP WHICH IS MORE ADVANCED THAN LINUX AND UBUNTU
[03:55] <Hobbsee> poopcheese: then why are you here?
[03:56] <ajmitch> to provide constructive criticism, obviously
[03:56] <Hobbsee> ajmitch: you haven't seen the half of it...
[03:56] <wgrant> Because with Windows you can entirely avoid .NET. Mhm.
[03:56] <ajmitch> I don't think I want to :)
[03:57] <wgrant> Anyway, that was some nice pre-lunch entertainment. Have fun!
[03:57] <Hobbsee> +z is a lovely mode...
[03:57] <ion_> hobbsee: You really want to still see his messages? :-)
[03:57] <Hobbsee> ion_: i'll deop soon.  or boot him.
[03:58] <Hobbsee> ion_: but sure.  he might actually become constructive.
[03:58] <Hobbsee> well, that was interesting.
[03:59] <StevenK> It becomes no fun if he can't troll.
[03:59] <Hobbsee> I think we need a sign saying "you need this much IQ to join this channel", though
[03:59] <vorian> i'll have to leave then :'(
[04:00] <Hobbsee> vorian: you can speak english, without capslock.    you're ok...
[04:00] <vorian> yay!
[04:03]  * Hobbsee ponders the ethics of forwarding to ##psychiatric-help-required or something
[04:14] <StevenK> slangasek: Are you around?
[04:16] <StevenK> slangasek: mysql-dsfg-5.0 failed to build on both i386 and amd64, I'm pondering just giving them back since the testsuite errors look ... odd
[04:19] <calc> is there a way to get ppa builds for port systems?
[04:19] <calc> eg easy testing of all archs without uploading to the main archive?
[04:24] <LaserJock> PPAs are Xen, so it's limited to what Xen can do, from what I understand
[04:24] <LaserJock> and so far that i386 and amd64
[04:25] <StevenK> And lpia
[04:26] <LaserJock> ah, right
[05:29] <NCommander> meh, launchpad is slow tonight
[06:41] <dholbach> good morning
[06:49] <NCommander> StevenK, where's the build log?
[07:44] <dholbach> can somebody moderate my mail on ubuntu-devel-announce?
[07:47] <NCommander> is kubuntu-desktop broken for anyone else?
[07:47] <cjwatson> dholbach: done
[07:47] <dholbach> cjwatson: thanks a lot
[07:47] <dholbach> cjwatson: you're up very early
[07:47] <cjwatson> NCommander: yes. http://people.ubuntu.com/~ubuntu-archive/testing/intrepid_probs.html
[07:48] <cjwatson> dholbach: couldn't sleep; about to go back and try again
[07:48]  * dholbach hugs cjwatson
[07:48] <NCommander> oh yay, the output of britney
[07:48] <NCommander> it looks like a few FTBFS are causing amd64 some grief
[07:50] <LaserJock> argg, a lot of edubuntu on that list :(
[07:50] <cjwatson> LaserJock: probably just due to kde
[07:50] <LaserJock> perhaps, I tried to clean up a bunch of that
[07:50] <NCommander> I didn't know Ebuntu was Kubuntu based
[07:51] <NCommander> It looks like python-qt4 is the holdup
[07:51] <cjwatson> it has some KDE components
[07:51] <LaserJock> I'll maybe get a chance to look at it more tomorrow
[07:51] <LaserJock> NCommander: it's Edubuntu not Ebuntu and it includes KDE-Edu
[07:51] <NCommander> er, my bad
[07:51] <NCommander> sorry, sorta late
[07:51] <NCommander> It looks like python-qt4 failed
[07:51] <NCommander> and then a bunch of packages got stuck behind it
[07:52] <LaserJock> we also have an edubuntu-desktop-kde meta package for KDE'ers
[07:52] <NCommander> edubuntu is susposed to be deployable over a thin client, right?
[07:53] <LaserJock> that's a common use-case yes
[07:55] <NCommander> I dunno, but GNOME seems kinda heavy to deploy to thin-clients, unless I'm mistaken that edubuntu is gnome based
[07:55] <LaserJock> well, it entirely depends on your environment
[07:56] <NCommander> well, whats the best way I could get started to try it out?
[07:56] <LaserJock> for sure some people with not-so-great-networks/servers or with a lot of clients find it better to go with something lighter
[07:56] <LaserJock> we used to ship XFCE as well, but when it got moved to Universe we stopped
[07:58] <LaserJock> but people can install whatever DE they like with the thin clients, the Edubuntu addon CD just provides an educational layer on top
[07:58] <LaserJock> NCommander: trying out what?
[07:58] <NCommander> edubuntu
[07:59] <NCommander> I dunno if I'm explaining myself well
[07:59] <LaserJock> well, you can grab an .iso, burn it, and pop it in :-)
[07:59] <NCommander> Why was Xfce moved to universe?
[07:59] <LaserJock> well, it's easier for people to work on it
[08:00] <NCommander> I thought canonical offered support for Xubuntu, or did that cease at some point
[08:00] <LaserJock> they never did
[08:00] <LaserJock> it used to be to build a CD using the Ubuntu infrastructure it had to be in Main
[08:00] <NCommander> Right. but they were able to fix that
[08:01] <LaserJock> but that's no longer the case, both Xubuntu and Ubuntu Studio build from Universe
[08:01] <NCommander> Is Edubuntu supported by Canonical?
[08:01] <LaserJock> yes
[08:02] <highvoltage> NCommander: Edubuntu has an IRC channel as well, on #edubuntu
[08:02] <NCommander> Neat
[08:02] <NCommander> I'm suprised though no other *buntu distros have shot up
[08:02] <LaserJock> how do you mean?
[08:02] <LaserJock> there's a bazillion of them on distrowatch ;-)
[08:02] <NCommander> I know emgent was trying to get wmaker-ubuntu off the ground, and Fluxbuntu also been working
[08:02] <NCommander> I mean offically on launchpad
[08:03] <LaserJock> well, it's rather difficult to build a derivative within the archive you're deriving from :-)
[08:03] <LaserJock> not impossible by any means, but the "easy" way is to fork off your own repo
[08:04] <NCommander> Well, I think emgent was simply going to try and get a meta package into universe, then use the cd burner scripts
[08:04] <LaserJock> mhm
[08:04] <NCommander> which is how I would roll any new distribution
[08:04] <NCommander> Unless for some reason I needed packages that were unsuitable for the *verse
[08:04] <LaserJock> I know of at least 5 derivatives in Universe
[08:05] <NCommander> Myth, Xfce, Studio
[08:05] <NCommander> What are the other two?
[08:05] <LaserJock> Ichthux and Ubuntu ME
[08:05] <raphink> hello
[08:05] <LaserJock> heh, hi raphink
[08:06] <raphink> hi LaserJock :)
[08:06] <NCommander> I didn't realize Ubuntu ME was that far off the ground yet, I thought it was still in conceptional stages.
[08:06] <jpds> hi raphink
[08:06] <raphink> hello jpds
[08:07] <NCommander> morning jpds
[08:09] <NCommander> LaserJock, once kfreebsd-i386 gets off the ground, I won't mind seeing an Ubuntu distribution ported to it (I've been keeping my eye on nexenta, but it seems to be struggling)
[08:11] <raphink> what does nexenta have to do with kfreebsd ?
[08:11] <raphink> just wondering
[08:12] <NCommander> I don't like the Linux kernel
[08:12] <LaserJock> Linux without the Linux ;-)
[08:12] <NCommander> I actually have a strong dislike for it
[08:12] <raphink> hehe I see ;)
[08:12] <NCommander> (I ran NetBSD for years before I finally got sick of ports)
[08:13] <raphink> and darwing is dead, too
[08:13] <raphink> although you can still use macports
[08:13] <NCommander> fink is alive and kicking
[08:13] <raphink> yes
[08:13] <NCommander> I actually learned a LOT abotu Debian packaging through fink
[08:13] <raphink> ah ok
[08:13] <NCommander> Mostly because I helped debug a rather nasty bug four or five years ago
[08:13] <NCommander> And helped bring the i386-darwin port into the world before Intel OSX  was released :-)
[08:14] <raphink> I use MacOSX, but I strongly dislike how I can't get my hands behind the polished stuff
[08:14] <NCommander> I won't mind seeing a Debian/Darwin
[08:14] <raphink> at least not as easily as on linux
[08:14] <NCommander> What I want is a kfreebsd designed system (jails :-)).
[08:14] <raphink> ok
[08:14] <NCommander> Solaris just ... well, it plain sucks on laptops
[08:14] <raphink> but it rocks on servers
[08:15] <NCommander> yeah
[08:15] <NCommander> my server is running kfreebsd-i386
[08:15] <NCommander> Maybe if I get bored this weekend I'll bootstrap Xubuntu kfreebsd-i386
[08:15] <NCommander> :-)
[08:15]  * NCommander is smacked a few times with a tree
[08:16] <LaserJock> a Linux tree?
[08:16] <LaserJock> :-)
[08:16] <NCommander> The freebsd kernel is actually available in Ubuntu
[08:16] <NCommander> apt-get source freebsd
[08:33] <StevenK> NCommander: http://launchpadlibrarian.net/17093486/buildlog_ubuntu-intrepid-i386.mysql-dfsg-5.0_5.0.67-0ubuntu3_FAILEDTOBUILD.txt.gz
[08:34] <mdke> with Ubuntu 8.04.1 my computer keeps freezing so that the only solution is to turn it off or reset it. How does one go about figuring out what the cause of that is and collecting useful information to file a bug?
[08:34] <mdke> I have a hunch that it's firefox because it always seems to happen when I am using firefox, but that might just be because I use firefox more than other applications
[08:35] <NCommander> StevenK, looks like the buildd had a seizure somewhere in there
[08:36] <StevenK> NCommander: I'm pondering bashing the give back button
[08:36] <NCommander> what's stopping you?
[08:36] <StevenK> Fear of breakage
[08:37] <lucas> mdke: check your ram?
[08:37] <NCommander> StevenK, how can be be broken worse?
[08:37] <NCommander> StevenK, did it FTBFS with the same error on multiple architectures?
[08:37] <StevenK> NCommander: Similar error
[08:38] <StevenK> NCommander: IE, test suite failure on amd64
[08:38] <StevenK> However, i386 is a little more important since it builds Arch: all
[08:38] <NCommander> is it a new upstream release or whatnot?
[08:38] <StevenK> NCommander: Nope
[08:38] <NCommander> what was the change?
[08:39] <mdke> lucas: ah, I'll try that, thanks
[08:39] <mdke> lucas: I hadn't thought that it could be hardware related
[09:34] <StevenK> NCommander: Damn it, mysql failed the same way
[09:34] <NCommander> StevenK, I'm looking at it
[09:34] <NCommander> I can't figure out what changed
[09:34] <StevenK> NCommander: I wonder if you have a machine faster than me to build it on
[09:35]  * NCommander grabs the soruce
[09:35] <NCommander> This is the version in intrepid, right?
[09:35] <StevenK> NCommander: Right
[09:35] <NCommander> Ok, I'll see if I can get it to build
[09:35] <StevenK> NCommander: palmer is no slouch, and it takes it nearly an hour to build
[09:35] <NCommander> Well
[09:35] <NCommander> I have a distcc cluster
[09:36] <NCommander> -j10 ;-)
[09:36] <StevenK> NCommander: It gives a log, I'd like to see the log
[09:36] <NCommander> The testing log, right?
[09:36] <NCommander> (I wasn't going to roll it in a pbuilder instance yet, so I can do some debugging)
[09:37] <NCommander> Its almost as if your causing the connection thread to crash and burn
[09:37]  * NCommander looks at the changelog
[09:37] <NCommander> It looks like the PIE hardening is what broke it
[09:37] <StevenK> Yes
[09:38] <NCommander> bah, I use dto run down this issue with m68k
[09:38] <soren> Wow. A package that manages to build on HPPA, but fails on i386 and amd64.. That's a first (to me).
[09:38] <StevenK> soren: Yes.
[09:38] <NCommander> Ouch
[09:38] <NCommander> Given the sorry state of hppa's compiler
[09:38] <StevenK> What is this PIE hardening, and why did it break MySQL?
[09:38] <NCommander> PIE stands for position independent code/executable
[09:38] <NCommander> StevenK, ever program in ASM?
[09:39] <StevenK> Oh, right
[09:39] <NCommander> Its got a performance hit on i396
[09:39] <NCommander> *i386
[09:39] <StevenK> Now I get it, I just hadn't heard of it in terms of hardening
[09:39] <NCommander> Yeah, that's a first for me
[09:39] <NCommander> I mean, I get the theory
[09:40] <NCommander> But Windows is completely position dependent
[09:40] <NCommander> I"ve never heard as PIE as a hardening defense
[09:40] <StevenK> Why do I have this feeling that it's SEGVing or something when it runs that test
[09:40] <NCommander> StevenK, its probably a thread reaching an abend
[09:40] <NCommander> But its got to be a compiler issue
[09:40] <NCommander> PowerPC is PIE by default (amd64 too), and they're not blowing up
[09:40] <NCommander> ER wait
[09:40] <NCommander> Stupid quesiton
[09:40] <NCommander> Is this failing on just i386?
[09:41] <NCommander> Cause I'm compiling on amd64 ;-)
[09:41] <StevenK> amd64 blew up too
[09:41] <NCommander> Ok
[09:41] <NCommander> good
[09:41] <soren> What the...
[09:41] <NCommander> BTW, mind doing me a favor as I hunt this down?
[09:41] <soren> http://launchpadlibrarian.net/17103261/buildlog_ubuntu-intrepid-hppa.mysql-dfsg-5.0_5.0.67-0ubuntu3_FULLYBUILT.txt.gz
[09:41] <NCommander> Can you sponsor an upload into Debian?
[09:41] <soren> The hppa builds shows lots of test failures, too, but doesn't seem to care.
[09:41] <NCommander> WTF
[09:41] <NCommander> Well
[09:41] <NCommander> UGH
[09:41] <NCommander> MY EYES
[09:41] <StevenK> Haha
[09:41] <StevenK> NCommander: I suppose
[09:42]  * Hobbsee notes this has nothing to do with her myspace page, either.
[09:42] <NCommander> I'm an m68k buildd admin
[09:42] <NCommander> I've sen funky failures
[09:42] <NCommander> Perl 5.10 and linuxthreads
[09:42] <NCommander> Yum
[09:42]  * StevenK tells his Sid chroots, "Update thyselves!"
[09:43] <StevenK> Hobbsee: Twitch
[09:43]  * Hobbsee grins
[09:43] <NCommander> StevenK, http://mentors.debian.net/debian/pool/main/c/codeblocks/codeblocks_8.02-1.dsc
[09:43] <Hobbsee> StevenK: it's *lovely*!
[09:43] <NCommander> As an aside, slangasek already took off running
[09:43] <NCommander> Upstream shipping a debian folder
[09:43] <NCommander> *shipped
[09:43]  * StevenK joins his fellow Steve in running away
[09:43] <NCommander> Crap
[09:43] <NCommander> Lost another DD
[09:43] <StevenK> :-P
[09:44] <NCommander> seb128 was going to do it, but his sid chroot "mysteriously" broke
[09:44] <StevenK> Sounds ominious
[09:44]  * NCommander watches StevenK randomyl break
[09:44] <NCommander> Seriously, how the hell does the hppa buildd not FTBFS
[09:44] <NCommander> That just plain hurts
[09:44] <StevenK> Haha
[09:45]  * NCommander looks at the harding patch
[09:45] <NCommander> I swear
[09:45] <NCommander> if this is because some idiot doesn't know how to apply -PIE, I'm going to hurt the security team
[09:45] <StevenK> NCommander: -fPIE?
[09:46] <NCommander> Well, a good source of segfaults is when -fPIE and -fnotPIE combine
[09:46] <NCommander> By your powers combined, I am caption Segfault!
[09:46] <NCommander> I ... WTF
[09:46] <StevenK> Yes. By your powers combined, I'm Captain SIGFPE
[09:47] <NCommander> The HPPA build did fail its test suite
[09:47] <NCommander> Miserably
[09:47] <StevenK> Oh. Dear. God. We channeled the same thing.
[09:47] <StevenK> Doing. It. Wrong.
[09:47] <NCommander> The servers were restarted 122 times
[09:47] <NCommander> Hrm
[09:47] <NCommander> Oh shit
[09:47] <NCommander> I see the issue
[09:47] <NCommander> I'm going to hurt the security team if I'm right
[09:48] <StevenK> No fair huring kees.
[09:48] <StevenK> Er, hurting
[09:48] <NCommander> Er, ok
[09:48] <NCommander> I see why it failed on hppa I think
[09:48] <NCommander> Was the the same failure on amd64?
[09:49] <StevenK> i386 failed in the subselect test
[09:49] <StevenK> amd64 failed in ps_7ndb
[09:49] <NCommander> it looks like its trying to call code from outside mysql that isn't PIE
[09:50] <NCommander> Which would cause an instant segfault
[09:50] <StevenK> Why would ps_7ndb pass on i386, and fail on amd64? :-)
[09:50] <StevenK> And subselect passes on amd64 and fails on i386
[09:51] <NCommander> StevenK, its running the test suite
[09:51] <StevenK> I think the MySQL test suite needs more kittens
[09:51] <NCommander> We could simply back out the hardening_patch
[09:51] <NCommander> binlog_killed_simulate         [ skipped ]   Test need debug binaries
[09:51] <NCommander> Bah
[09:51] <NCommander> Fixing this might be a plain nightmare
[09:53] <StevenK> NCommander: In terms of with PIE-hardening?
[09:53] <NCommander> Probably
[09:53] <NCommander> ever look at mysql's codebase?
[09:54] <StevenK> Nope. And I'd rather keep my eyes
[09:54]  * NCommander codes in cobol for fun sometimes
[09:54] <NCommander> I perfer my sanity :-)
[09:54] <StevenK> "Fun" you say
[09:54] <NCommander> I also like x86 ASM
[09:54] <NCommander> What messes with my head
[09:55] <NCommander> Is I've seen x86 ASM thats more clear to me then your average perl script
[09:55] <StevenK> My last job was as a Perl coder
[09:55]  * StevenK knows more about Perl internals than he cares to admit
[09:55] <NCommander> "You shoot yourself in the foot, but six months later, you have no idea how you did it"
[09:56] <NCommander> I had to fix a FTBFS deep within herny on m68k
[09:56] <NCommander> That was wonderful
[09:56] <NCommander> mostly because building perl was something like a 10 hour job
[09:56] <StevenK> If you use a doorstep arch, it is
[09:56]  * StevenK hides
[09:56]  * NCommander crowbars StevenK 
[09:57] <StevenK> :-P
[09:57] <NCommander> Hurd is a doorstep arch, we actually *gasp* released a few times
[09:57]  * wgrant clobbers StevenK with an m68k box.
[09:57] <NCommander> wgrant, use an s390 one
[09:57] <NCommander> Leaves a bigger mark
[09:57]  * StevenK counters with the ultrasparc he has lying around here
[09:57]  * NCommander uses slangasek's alpha
[09:57] <NCommander> BHAHAHAHAHA
[09:58]  * StevenK has one of those, too
[09:58] <NCommander> you know
[09:58] <NCommander> if mysql passes its test suite
[09:58] <NCommander> I'm going to be pissed as hell
[09:58] <NCommander> I hate migrating bugs
[09:58] <StevenK> What did you change?
[09:58] <NCommander> nothing
[09:58] <NCommander> But I have that funny feeling
[09:58] <NCommander> that the PIE bug will appear differently on different machines
[09:59] <NCommander> Ugh
[09:59] <NCommander> ndb_alter_table                [ disabled ]  failed for some reason
[09:59] <NCommander> O_O;
[09:59] <StevenK> Haha
[09:59] <NCommander> Well, now I know why hppa didn't actually "fail"
[09:59] <StevenK> No, just 3/4 of the test suite was disabled
[09:59] <soren> NCommander: I'm quite sure kees did a test build with PIE enabled before he uploaded it. Has anyone talked to him?
[10:00] <NCommander> Well, I'm getting the feeling looking at the build logs we're dealing with a migrating bug
[10:00] <NCommander> i.e., it might appear different places based on kernel version, compiler version, etc.
[10:00] <StevenK> soren: That's why I'm retiscent to just upload a fix. It's 2am in $TZ, though
[10:00] <NCommander> The quick fix is to removing the harding-wrapper
[10:00] <NCommander> And let security figure out the FTBFS
[10:01] <NCommander> But given the fact the test suite was timing out with -fPIE enabled
[10:01] <NCommander> I dunno
[10:01] <NCommander> That sounds like quite a performance hit
[10:01] <NCommander> and at best, a minor security fix
[10:02] <NCommander> and only really help if your using something that randomizes the address space to prevent a return-to-libc attack
[10:02] <StevenK> I thought we already enabled address space randomization in Hardy
[10:04] <NCommander> I can't say I know
[10:05] <NCommander> But its ineffective expect on AMD64
[10:05] <NCommander> Its possible to brute force a return-to-libc attack on i386
[10:05] <NCommander> Its kinda said mysql built on every architecture expect the release ones -_-;
[10:06] <NCommander> wait
[10:06] <NCommander> ...
[10:06] <NCommander> WTF?!
[10:06] <NCommander> #   intrepid lpia   Successfully built  (DONE)
[10:06] <NCommander> HOW IS THAT IN THE LOVE OF GOD POSSIBLE
[10:06] <StevenK> Yes.
[10:06] <NCommander> o_o;
[10:06] <NCommander> I.
[10:06] <NCommander> O_o;
[10:06]  * NCommander is speechless
[10:07] <NCommander> lpia doesn't include any glibc/compiler changes does it?
[10:07] <StevenK> Um, just compiler flags
[10:08] <NCommander> YEah
[10:08] <NCommander> It looks like the issue on hppa/amd64 is the ndb storage library
[10:08] <NCommander> i386 is suffering from its own issues
[10:11] <NCommander> StevenK, your going to love this
[10:11] <NCommander> ps_7ndb                        [ pass ]          10331
[10:11] <NCommander> (on amd64)
[10:13] <NCommander> StevenK, were you able to reproduce the FTBFS on 127.0.0.1?
[10:14] <StevenK> NCommander: I didn't try it
[10:14] <NCommander> StevenK, Ok, whoever did the harding seems to not have read the MySQL manual
[10:15] <NCommander> It needs --with-pic passed as a configure script option
[10:15] <NCommander> Or things will explode it seems
[10:16] <StevenK> NCommander: Ahhh
[10:16] <NCommander> I'm not sure if it will make a difference yet
[10:17] <NCommander> Note
[10:17] <NCommander> Oh
[10:17] <NCommander> That's pretty
[10:17] <NCommander> -fPIC is passed as an arguement already
[10:18] <NCommander> Now
[10:18] <NCommander> If memory serves
[10:18] <NCommander> If the same option is passed to GCC more than once
[10:18] <NCommander> Doesn't it act like a toggle?
[10:20] <StevenK> NCommander: Manual page doesn't say so
[10:20] <NCommander> Hrm
[10:21] <NCommander> It's stack-protector
[10:21] <NCommander> Known issue with mysql and -fstack-protector
[10:21] <NCommander> IT causes random segfaults
[10:21] <NCommander> BTW
[10:21] <NCommander> All 492 tests were successful.
[10:21] <NCommander> :-P
[10:22] <NCommander> so, to fix this
[10:22] <NCommander> We need to disable the stack protector
[10:24] <NCommander> StevenK, I'm rolling a patch
[10:24] <StevenK> I thought the stack protector was enabled in Hardy?
[10:24] <NCommander> Programs won't use it if not compiled with -fstack-protector
[10:25] <wgrant> Stack protector has been on by default since before Hardy, IIRC.
[10:25] <wgrant> It is at least in Hardy.
[10:25] <NCommander> Meh
[10:25]  * NCommander scratch that
[10:25] <NCommander> My mistake
[10:25] <NCommander> Unless mysql turns it off
[10:26] <NCommander> Well, solving this now is a process of elimation
[10:27] <NCommander> If *ahem* I could reproduce the FTBFS
[10:27] <soren> wgrant: Since edgy.
[10:28] <wgrant> soren: I thought so, but I don't recall specifically verifying that before Hardy. Thanks.
[10:28] <soren> NCommander: GCC in Ubuntu defaults to -fstack-protector. You need to explicitly pass -fno-stack-protector to disable it.
[10:28] <NCommander> Ok
[10:28] <NCommander> So that rules that out
[10:28] <NCommander> I'm looking for known bugs
[10:29] <NCommander> It might be worth turning off the hardening PIE, and let mysql handle it directly
[10:29] <NCommander> But ... bah
[10:29] <NCommander> Until I can recreate this on demand ...
[10:31] <StevenK> And pass --with-pic ?
[10:31] <NCommander> StevenK, maybe
[10:31] <NCommander> I'm uploading the package as is to my PPA
[10:31] <NCommander> I need to confirm the segfaults
[10:31] <NCommander> Since it passed my system without segfaulting
[10:32] <NCommander> and we know the issue is intermittiant
[10:32] <NCommander> lpia passed, while i386 didn't
[10:32] <NCommander> And palmer failed, but mine passed
[10:33] <StevenK> NCommander: palmer won't build for the PPA, either
[10:33] <NCommander> I'm aware of that
[10:33] <NCommander> But if I can't reproduce here, then I got to try reproducing in the PPA
[10:33] <NCommander> Unless you want to start randomly feeding the buildds packages and see what builds
[10:34] <StevenK> NCommander: Well, sure.
[10:34] <StevenK> NCommander: Merely stating
[10:35] <StevenK> Aww. ~ncommander doesn't exist
[10:35] <NCommander> Nope
[10:35] <NCommander> Its ~sonicmctails
[10:35] <NCommander> :-)
[10:36] <NCommander> https://edge.launchpad.net/ubuntu/+ppas - BTW, I dunno if you saw this, but needless to say, I have no life :-)
[10:36] <NCommander> 55 uploads by hand in a week
[10:37] <soren> NCommander: Yes, please calm down. You making everyone else look bad.
[10:37] <NCommander> Well, someone asked me to package xfce 4.4, and then gnat 4.2
[10:38] <NCommander> It's not my fault there are a lot of things to build :-P
[10:38] <StevenK> NCommander: Doing NBS work for Gutsy I uploaded ~ 95 packages in the space of 3 hours
[10:38] <NCommander> NBS?
[10:38] <StevenK> Not Built from Source
[10:38] <NCommander> Right, but uploaded?
[10:38] <StevenK> Remember the libcurl mess?
[10:38] <NCommander> Vaguely
[10:39] <NCommander> Didn't the soname not change on an ABI break?
[10:39] <StevenK> NCommander: That was ~ 95 uploads, most of them no-change rebuilds (-4 -> 4build1, for example), but I still added changelog entries, rebuilt the source, signed ~ 90 changes files and uploaded the whole lot.
[10:40] <NCommander> my pipe couldn't handle that many uploads in such a short peroid of time
[10:40] <StevenK> NCommander: Only source, and no origs
[10:40] <NCommander> Oh
[10:40] <NCommander> n/m ;-)
[10:40] <NCommander> I didn't realize you joined my cruft busting group
[10:40] <NCommander> (I need to send some emails to it, but I keep getting sidetracked)
[10:40] <StevenK> I've been cruft busting since Edgy
[10:41] <NCommander> Sweet
[10:41]  * NCommander bends buildd.py to work on PPAs
[10:41] <NCommander> I've discovered how anonying poor buildd administration can be
[10:42] <NCommander> Especially if you have seven or eight packages pending builds, and they form a chain of dependencies
[10:42] <jpds> NCommander: If you can make a --ppa flag for it, please give me a patch against lp:ubuntu-dev-tools
[10:42] <StevenK> NCommander: If they build depend properly, most of them should hit DEPWAIT and then retry themselves
[10:43] <NCommander> I've never seen a PPA package properly leave dep-wait
[10:43] <NCommander> (they enter it just fine, just never leave)
[10:44] <StevenK> Hmm. Hotel California buildd states
[10:44] <NCommander> Yeah, but they haven't been like that since '79
[10:44] <NCommander> StevenK, you can help speed up the process by rescoring my PPA to take priority on the buildds ;-) *smacked*
[10:45] <NCommander> (it sucks when the i386 translations build hits and i386 builds take forever)
[10:46] <NCommander> StevenK, this is going to take awhile to build, got any other strange FTBFS that need looking at?
[10:46] <Fritz> anyone willing to help with a question?
[10:46] <StevenK> NCommander: That's it
[10:46] <NCommander> Do I need to get this done before the FF?
[10:46] <NCommander> Or can we get an FFE for it
[10:47] <StevenK> NCommander: It's fixing a build failure ...
[10:47] <NCommander> so FFE :-)
[10:47] <NCommander> Just making sure
[10:47] <NCommander> I work well under pressure
[10:47] <NCommander> SO I wanted to know how fast I had to turn around a patch
[10:48] <StevenK> NCommander: In which case, I wanted the patch yesterday.
[10:48] <NCommander> This will probably take a day or two to run down. I'm seeing if I can reproduce on real i386 hardware
[10:48] <NCommander> (as in hardware I control)
[10:49] <Fritz> i have an hp tx1000, wifi and touch screen funtions are not working with ubuntu, any ideas?
[10:49]  * StevenK peers at samarium
[10:49] <NCommander> Fritz, wifi probably means firmware missing
[10:49] <Fritz> what can i do?
[10:49] <NCommander> Check Restricted Drivers (System -> Administration -> Restricted Drivers)
[10:49] <NCommander> See if one is available for your wifi card
[10:50] <NCommander> Note: You need to be connected to the internet on that machine to get the firmware
[10:50] <Fritz> ethernet wont work, can i do it comp to comp, or use windows drivers?
[10:51] <NCommander> If you have windows drivers
[10:51] <NCommander> ndiswrapper might get you online
[10:51] <NCommander> But I'm sorta blocking on the specifics
[10:51] <NCommander> First see if Ubuntu offers to download firmware
[10:51] <Fritz> will do, thank you.
[10:52] <NCommander> If it does, then your at least in the right ball part w.r.t.  to getting wifi going
[10:52] <NCommander> s/part/park/g
[10:52] <NCommander> Man, my list of icons has gotten very long O_O;
[10:54] <NCommander> StevenK, why were you looking for my profile page out of curosity
[10:57] <StevenK> NCommander: To find your PPA
[10:58] <NCommander> Ah, "You are at a long and slow Launchpad. Exits are in all directions"
[10:58] <NCommander> ^Launchpad page
[10:59] <NCommander> StevenK, BTW, if this doesn't FTBFS
[10:59] <NCommander> I think I'm going to loose my mind
[11:03] <cjwatson> you should always loose your mind anyway; keeps it free from unnecessary constraints ...

[11:03] <StevenK> Bwaha
[11:05] <NCommander> Prolonged exposure to debian-cd already loosed my grip on my sanity
[11:05] <NCommander> lol, I say debian-cd, and he drops
[11:06] <NCommander> StevenK, so if it passes in the PPA, then what?
[11:06] <NCommander> I can't reproduce, and its so hard to reproduce, its hard to tell if the harden-security is a regression or not
[11:06] <StevenK> Then I'm not sure.
[11:06] <StevenK> Then I think we beg cjwatson
[11:06] <NCommander> for all we know, this issue been going on since 0ubuntu1
[11:06] <NCommander> cjwatson, please build xubuntu powerpc CDs :-)
[11:06] <NCommander> Oh wait
[11:07] <cjwatson> beg me for what?
[11:07] <NCommander> wrong begging
[11:07] <StevenK> Nah, 0ubuntu2 added the patch
[11:07] <NCommander> No, I mean before hardened was enabled
[11:07] <NCommander> Its possible this is a lurking issue
[11:07] <NCommander> cjwatson, we're dealing with an FTBFS on amd64/i386, but passes on lpia/hppa :-)
[11:07] <StevenK> cjwatson: mysql-dfsg-5.0 FTBFS on i386 and amd64 due to an unclear issue
[11:07]  * NCommander hears the scream from here
[11:07] <NCommander> sweet, sweet music
[11:08] <NCommander> Maybe it means palmer having hardware issues
[11:08] <StevenK> I hope not.
[11:09] <NCommander> actually, it would also mean the amd64 buildd having them too
[11:09] <NCommander> DktrKranz, how goes the gnat-4.2 transition
[11:09] <cjwatson> so this is "random failure of user_limits.test
[11:09] <cjwatson> "?
[11:09] <NCommander> and how can I help to speed that up before the freeze
[11:09] <NCommander> cjwatson, no, its random tests on different goes
[11:09] <NCommander> It passed on my amd64
[11:09] <NCommander> hppa had a non-critical failure
[11:09] <NCommander> It failed on the buildd
[11:09] <NCommander> the security machine where they did the work likely passed it
[11:10] <DktrKranz> NCommander: it involves hardy, no need to hurry right now, better focus on pre-FF bugs now
[11:10] <NCommander> Ok
[11:10] <NCommander> I may move the packages into a temp group
[11:10] <cjwatson> NCommander: has anyone tried the upstream commit http://lists.mysql.com/commits/49751 (referenced from http://bugs.mysql.com/bug.php?id=23921)? that fix isn't specific to a single test
[11:10] <NCommander> my PPA getting cluttered
[11:10] <NCommander> cjwatson, No, but this might be the issue
[11:10] <NCommander> The PIC code slows down mysql slightly
[11:10] <NCommander> Might JUST be enough to trigger a random failure
[11:11] <NCommander> expect palmer made the same failures
[11:12] <NCommander> If this doesn't fail
[11:12] <NCommander> I'll backport this fix, and then hit StevenK to try it on the real buildds
[11:13] <NCommander> If THAT does fix it, I think we can rule it out to the hardening fixes (which is what I thought it was since the FTBFS only showed up after the security team added them)
[11:14] <NCommander> Is mysql main or universe?
[11:14] <NCommander> (if I don't have to constantly torment StevenK, thats a good thing for my future as a potential core-dev)
[11:16] <StevenK> Main
[11:16] <NCommander> so much for core-dev
[11:16] <NCommander> Although at this point, UUC would be nice :-)
[11:18] <NCommander> cjwatson, connection issues?
[11:19] <NCommander> StevenK, so how goes reviewing codeblocks ;-)
[11:21] <StevenK> I ran screaming, remember?
[11:21] <NCommander> crap
[11:21] <NCommander> Well, I can run screaming from this bug
[11:21] <NCommander> So run back
[11:22] <cjwatson> NCommander: router was playing up so I rebooted it
[11:22] <NCommander> My router runs Debian mips :-)
[11:24] <NCommander> so how goes your morning cjwatson
[11:25] <cjwatson> not inclined to random chat with FF tomorrow, really :)
[11:25] <NCommander> ah ok
[11:25] <NCommander> StevenK, why aren't you going crazy with the last day before FF?
[11:26] <NCommander> StevenK, it passed the build that failed on the main buildd
[11:26] <NCommander> s/build/test/g
[11:26] <StevenK> NCommander: I probably should be
[11:26] <NCommander> so what gives ;-)?
[11:27] <NCommander> (if you have things for me to do, I'd be glad to work while I wait for these things to build)
[11:27] <StevenK> NCommander: Currently off duty while waiting for dinner
[11:27] <StevenK> I'll panic when I'm working again
[11:27] <NCommander> Well, if you need some FTBFS/packaging/Ada work done
[11:27] <NCommander> Drop me an email
[11:28] <NCommander> (yes, I said ada)
[11:31] <NCommander> StevenK, you are going to love this
[11:32] <NCommander> It passed the test that failed
[11:38] <NCommander>  uh
[11:38] <NCommander> Shit
[11:38] <NCommander> I think I crahsed the PPA machine
[11:38] <NCommander> StevenK, FTBFS/packaging/Ada
[11:39] <NCommander> er
[11:39] <NCommander> StevenK, https://edge.launchpad.net/+builds
[11:47] <StevenK> NCommander: Hm, I see that.
[11:47] <StevenK> NCommander: "Bad"
[12:17] <jpds> cjwatson: Have you seen http://labs.mozilla.com/2008/08/introducing-ubiquity/ ?
[12:19] <cjwatson> jpds: heh. well, it's an English word, it's going to be hard to claim a monopoly
[12:23] <Treenaks> cjwatson: tell that to the terminator people :)
[12:31] <Ng> Treenaks: harsh ;)
[12:33] <Treenaks> Ng: well, I heard they're considering a name change because another project had the terminator name
[12:33] <Treenaks> Ng: same with firebird->firefox, but I'm not sure if that's a "proper" English word
[12:33] <Pici> terminatorx probably...
[12:33] <jpds> Treenaks: Ng is terminators author.
[12:33] <Treenaks> jpds: ah! didn't know
[12:33] <Treenaks> Ng: sorry then :)
[12:33]  * Pici either
[12:34] <cjwatson> I'm not renaming ubiquity again. The Mozilla project is different enough that, well, whatever.
[12:34] <Ng> Treenaks: we're considering it, but probably not doing it ;)
[12:35] <Treenaks> cjwatson: also, the Mozilla project has renamed before.. :)
[12:37] <thom> i'd think a bigger objection to terminator is that it's un-googleable :)
[12:39] <tacone> thom: "terminator linux"
[12:39] <Treenaks> terminator -skynet
[12:40] <tacone> terminator -hasta -la -vista -baby ?
[12:46] <Ng> thom: I think the kind of people likely to be using multiple terminals simultaneously can cope with that :)
[12:48] <davmor2> Guys why do any graphical changes you make in OEM mode not get carried forward to end user?  E.g. change backdrop, switch on compiz etc
[12:50] <cjwatson> because those are stored in the home directory, and it's extremely unclear what if anything should be copied over from there to the newly created user
[12:50] <cjwatson> for example, it's not unusual for the username to end up hardcoded in files in the home directory (e.g. absolute paths)
[12:50] <cjwatson> the safe option is to create the new user from scratch, and that's what we do
[12:51] <cjwatson> $ grep -r cjwatson .gconf | wc -l
[12:51] <cjwatson> 12
[12:53] <davmor2> cjwatson: doesn't that then negate the whole purpose of the oem mode?  Isn't the idea that you set things up as you want the end user to see them and then the end user adds there details?
[12:54] <cjwatson> you set the computer up, not necessarily the user account
[12:54] <cjwatson> if you want to set up default user accounts, use something like sabayon
[12:54] <cjwatson> s/default /defaults for /
[12:55] <davmor2> OKay thanks for that then :)
[12:56] <cjwatson> the core purpose of the oem mode is to allow the end user to enter their personal settings the first time they boot a system that somebody else installed
[12:56] <cjwatson> the customisation possibilities afforded by that kind of arrangement are incidental, not the whole purpose
[13:06] <StevenK> cjwatson: Okay, can we talk about this MySQL thing? Should I prepare an upload with that patch, since NCommander's test on the PPA buildds has caused samarium to offline itself twice now.
[13:08] <cjwatson> StevenK: my direct knowledge of MySQL can be counted on the fingers of one foot; if you think it's sane, you should probably go ahead
[13:09] <StevenK> cjwatson: Same. But MySQL being broken is causing ubuntu-mid to be uninstallable, which is holding me up, so I care until it's fixed, either by me or someone else.
[13:09] <cjwatson> I'd JFDI if I were you
[13:09] <Hobbsee> cjwatson: he'll quote you on that..
[13:09]  * StevenK grins
[13:10] <StevenK> "But Colin said yes!"
[13:10]  * ogra wonders what in -mid depends on mysql
[13:14] <StevenK> ogra: gstreamer-plugins, somehow
[13:14] <ogra> bah
[13:15] <StevenK> Yes. I don't get it either.
[13:16] <ogra> StevenK, hmmm, i see gstreamer0.10-plugins-bad and gstreamer0.10-plugins-ugly in the -mid deps
[13:17] <StevenK> ogra: That's what I meant
[13:17] <ogra> we should only have -base and -good
[13:27] <ogra> kirkland, ping (i was contacted by edubuntu users about #120375 ... you need an upload bitch ?)
[13:28] <zul> ogra: kirkland is still on holidays but he pops in from time to time
[13:28] <ogra> zul, ok
[13:28] <ogra> zul, thanks :)
[13:28] <ogra> he at least leaves his clientconnected, so i count on the fact that he reads the ping at some point :) )
[13:32] <zul> ogra: yep he does he is backpacking in scotland though apparently
[13:39] <Riddell> hmm, mvo on holiday, guess I'll just have to hijack UpdateManager without permission
[13:43] <cjwatson> tseliot: can we make the nvidia modalias packages architecture: all? the current situation causes uninstallability on architectures other than amd64 and i386
[13:44] <tseliot> cjwatson: but AFAIK NVIDIA doesn't support other architectures or am I missing something?
[13:45] <cjwatson> tseliot: well, in that case make nvidia-common architecture: amd64 and i386 only
[13:45] <cjwatson> can't have it both ways ;-)
[13:46] <cjwatson> tseliot: actually, though, I think you're mistaken. I know there have been powerpc systems shipped with nvidia cards.
[13:46] <tseliot> cjwatson: ah, ok I see your point now. I have to talk to mvo before I do it, since he uses nvidia-common in update manager
[13:46] <TheMuso> Yes powerpc systems did have nvidia cards.
[13:46] <cjwatson> Apple shipped its 12" PowerBooks with nvidia for some time
[13:46] <TheMuso> s/did/do/
[13:47] <TheMuso> and G5s as well.
[13:47] <\sh> grmpf....does anyone know how to force dkms and this somehow broken fglrx driver to build for intrepids kernel?
[13:47] <cjwatson> tseliot: making the modalias packages architecture: all shouldn't need mvo's help, I think. If you just need somebody to do the upload then I can do that
[13:47] <tseliot> cjwatson: ok but is there a proprietary driver (on Linux) which works on PPC?
[13:48] <tseliot> cjwatson: making it amd64 and i386 only might cause him some trouble though
[13:48] <TheMuso> tseliot: no there is not, and never likely to be.
[13:49] <tseliot> TheMuso: right that's my point.
[13:50] <tseliot> cjwatson: we should make nvidia-common amd64 and i386 only but I have to make sure that this doesn't cause problems to Update manager
[13:53] <jdstrand> soren: as you know, I have been working on ufw's package integration (see UbuntuFirewall). Basically, packages can declare profiles for ufw to use in its rules.
[13:54] <soren> jdstrand: Right.
[13:54] <jdstrand> soren: currently, packages need to drop files into /etc/ufw/applications.d, but I don't think I like that location, in part because if these changes ever get into debian, the location will likely need to be more general
[13:54] <jdstrand> eg /etc/fw.d
[13:54] <cjwatson> tseliot: I just checked: update-manager only build-deps on nvidia-common on amd64 and i386, and it takes care to cope gracefully if nvidia-common is missing at run-time
[13:55] <soren> jdstrand: You think?
[13:55] <jdstrand> soren: so that other applications could use them (the declaration themselves are not ufw speicific, just a title, description and port declaration
[13:55] <jdstrand> soren: :)
[13:55] <cjwatson> tseliot: so I don't think restricting nvidia-common's architectures should regress anything in update-manager
[13:55] <soren> jdstrand: Hm... I guess.
[13:55] <jdstrand> soren: so I was curious as to your opinion on this
[13:56] <jdstrand> cjwatson: if you'd like to weigh in, that would be great too ^^
[13:56] <tseliot> cjwatson: ok, then I'll change the package and ping you back with the link or would you prefer if I filed a bug report and put the link there?
[13:56] <cjwatson> tseliot: either works fine
[13:56] <tseliot> cjwatson: ok
[13:57] <jdstrand> soren: it isn't so much about other firewall apps using them, as much as making sure that packages don't have to change the location if another app wants to
[13:57] <jdstrand> use those profiles
[13:57] <soren> jdstrand: Right. Hmm..
[13:57] <zul> jdstrand: I take those ufw needs to be sponsored today?
[13:57] <cjwatson> (I suggest "firewall" rather than "fw", but that's trivial)
[13:57] <jdstrand> I kinda liked /etc/firewall.d, but that is used by cfengine (at least)
[13:58] <cjwatson> ah
[13:58] <cjwatson> I question the Hungarian .d
[13:58] <soren> O_o
[13:58] <ion_> /etc.d
[13:58] <jdstrand> zul: yes, but as I am changing a location, I will do all those
[13:58] <zul> jdstrand: coolio
[13:59] <jdstrand> cjwatson: I'm not married to it-- did it more out of habit
[13:59] <soren> cjwatson: Hungarian .d?
[13:59] <cjwatson> the original semantics of .d (as best as I can reverse-engineer them) were something that could also be in a parallel .conf file, but that had a split-out form
[13:59] <cjwatson> I don't think we should necessarily be using it for all directories :-)
[13:59] <jdstrand> heh
[13:59] <ion_> Indeed
[13:59] <cjwatson> soren: http://en.wikipedia.org/wiki/Hungarian_notation
[13:59] <jdstrand> actually, I wonder if firewall-profiles would be appropriate?
[14:00] <jdstrand> though that seems to imply too much
[14:00] <soren> cjwatson: Oh, right.
[14:00]  * soren totally didn't make that connection
[14:01] <cjwatson> jdstrand: do you think it's likely that such other applications will actually exist?
[14:01] <soren> And just as importantly: Do you think it'll share its config file format with ufw?
[14:01] <cjwatson> it's laudable to design for the future, but it's also easy (a) to go too far (b) to forget other things that such future applications will need and so wind up having to change lots of stuff anyway
[14:02] <cjwatson> there are packages in Debian that e.g. ship configuration for particular editors, and I think this is generally considered OK
[14:02] <jdstrand> cjwatson: doubtful, but I'd like to get ufw included in debian, and also the changes to the packages that drop files into this directory, so wanted a sane location
[14:03] <jdstrand> that said, they maintainer script would still need to call ufw, so Debian would need to accept ufw whole-hog anyway to accept those changes
[14:03] <cjwatson> I actually think it might be preferable to use your own namespace for now (and I don't think that would be as much of a problem for package integration in Debian as you think), rather than grabbing a nice chunk of namespace that (from Debian's perspective) you might not make good use of
[14:04] <jdstrand> which is more doubtful, so I suppose my current location is ok
[14:04] <cjwatson> why does the maintainer script need to call ufw, rather than ufw registering a trigger on /etc/ufw.d/ ?
[14:05] <jdstrand> cjwatson: TBH, I am not familiar with this process. can you point me somewhere to learn more?
[14:05] <cjwatson> I can point you to man-db as a (perhaps slightly overworked) example
[14:05] <cjwatson> the triggers specification is here: http://lists.debian.org/debian-dpkg/2007/04/msg00076.html
[14:05] <cjwatson> (but is very much package-manager-theoretical)
[14:06] <jdstrand> cjwatson: thanks. this is all excellent feedback
[14:06] <cjwatson> jdstrand: the menu package is a simpler example, I think
[14:07] <cjwatson> this works as long as all the maintainer scripts would just be calling ufw with the same arguments and having it import all the new configuration it finds
[14:07] <cjwatson> if they need to use different arguments it won't work as it stands
[14:08] <jdstrand> cjwatson: hmm, it needs one argument that is different
[14:08] <cjwatson> ah, in that case just do what you have to do for now and skip the triggers bit
[14:09] <cjwatson> however: what happens if ufw is installed *after* a package that ships a file in /etc/ufw/ ?
[14:09] <cjwatson> does it know how to import that configuration when it's configured itself?
[14:09] <jdstrand> cjwatson: yes
[14:10] <cjwatson> in principle, then, could you do whatever ufw.postinst does to import all configuration in every maintainer script?
[14:10] <tseliot> cjwatson: would this lintian error be tolerated? build-depends-indep-without-arch-indep
[14:11] <cjwatson> tseliot: no, that will cause problems - you need to change Build-Depends-Indep to Build-Depends in debian/control
[14:11] <tseliot> ok
[14:14] <jdstrand> cjwatson: the profiles in question aren't really imported per se. on each invocation, ufw checks the /etc/ufw/applications.d directory and uses what it finds. the postinst bits do two things-- update the running firewall if there are rules that reference the profile, and add a new rule for the specified profile if the admin setup ufw to do that
[14:14] <jdstrand> (by default ufw will not add any rules)
[14:15] <cjwatson> ok, ignoring my wording snafu though, do you see what I'm getting at?
[14:15] <TheMuso> For a RAID4 array, what is the minimum number of disks you need for the array to be functional, i.e not degraded?
[14:16] <pascal> Are others having trouble using the arrow keys in Intrepid? I installed Alpha 4 and after that I'm not able to use arrows keys, delete and some other keys. Works fine if I boot up in hardy
[14:16] <cjwatson> we're trying to use triggers where possible rather than adding a chunk to every maintainer script, because ultimately it's simpler to maintain, and it's generally faster (particularly when you take into account doing this for all the random things done in maintainer scripts)
[14:16] <jdstrand> cjwatson: IIUC, you're saying maybe I could make it triggerable
[14:16] <cjwatson> right
[14:17] <cjwatson> doing so usually involves having a way for the triggered package to just suck in all the stuff that was done, as if each package had run an individual command
[14:17] <cjwatson> for instance it's so much better for man-db to automatically update itself when packages update manual pages than it would be for every package to have to say mandb -f /usr/share/man/man1/ls.1.gz or whatever
[14:17] <jdstrand> cjwatson: right-- I really like the idea, and I can absolutely make it work with the 'update' part, but have to think about the '--add-new' part
[14:18] <cjwatson> ok, it shouldn't block your work for feature freeze or anything - I just wanted to raise it since it seemed like an obvious extension
[14:18] <jdstrand> (there was great interest in that aspect, though personally I don't like adding rules to a firewall via package installation)
[14:19] <tseliot> cjwatson: here's a file containing the links to the files: http://albertomilone.com/ubuntu/nvidia-common/links.txt
[14:20] <jdstrand> cjwatson: thanks again-- really excellent feedback :)
[14:23] <TheMuso> nvm got my answer
[14:27] <cjwatson> argh, dupload is consistently timing out while uploading ubuntu-meta (a whole 37KB)
[14:27] <ogra> still your network ?
[14:28] <cjwatson> I don't know what's wrong, scp is timing out on that file as well
[14:30] <\sh> bryce: this screen-resolution-extra stuff for the gnome applet, how does it work? I have here an ATI card, with radeon driver running, two screens, but I'm unable to move one screen to the left or to the right, but this applet says so. and nothing is changed in xorg ... how does it work then?
[14:30] <cjwatson> to multiple different hosts
[14:33] <cjwatson> weirdness threshold exceeded; I think I'll just reboot
[14:41] <ogra> tedg, seen Bug #229618 ?
[14:42] <tedg> ogra: Yeah, I'm curious what you're thinking there.  Merge the changelogs differently?
[14:42] <ogra> no idea, i'D just grab the mentioned 5.07 to get tormods changes in (i bet he has done them in debian)
[14:43] <ogra> beyond tht the usual megre policies apply
[14:44] <tedg> ogra: Sorry, I hadn't read the last comment...  now I'm up to speed.
[14:45] <tedg> My personal feeling is that tormod should do it, are version updates (5.05 to 5.07) effected by feature freeze?
[14:55] <\sh> grmpf...in gnome + kde4.1 the dual head stuff doesn't work with the radeon driver...regarding the logs it find everything and initializes it
[15:02] <ogra> tedg, yes, feature freeze == upstream version freeze ... though i do think getting an exception wouldnt be to hard, but i'm not the one to decide
[15:02] <ogra> \sh, dual head like one head gnome and the other kde ? :P
[15:03] <\sh> ogra: lol...
[15:03] <ogra> btw, sorry, but we didnt make it to froscon
[15:03] <\sh> ogra: na serious...I see the two screens...gnome: sending one screen to the left or right is dropping me to no signal on screens , and kde screen resolution tool doesn't do anything...
[15:04] <ogra> works fine here with itel card.... no radeon around ...
[15:04] <ogra> *intel
[15:06] <tedg> ogra: Okay, I'll reply to the bug and then do it if tormod doesn't have time.
[15:19]  * Adri2000 needs an archive admin
[15:25] <cjwatson> tseliot: done
[15:25] <cjwatson> ah, my networking problems were due to having both wired and wireless interfaces up at the same time
[15:25] <Adri2000> cjwatson: could you please copy two packages from -proposed to -updates for me?
[15:26] <tseliot> cjwatson: great :-)
[15:29] <cjwatson> Adri2000: bug number?
[15:30] <Adri2000> cjwatson: bug #252689
[15:30] <Adri2000> no, sorry
[15:30] <Adri2000> bug #243722
[15:30] <Adri2000> (was reading #launchpad at the same time :p)
[15:33] <cjwatson> Adri2000: ok, can't do it right this minute but I'll queue it up
[15:34] <Adri2000> ok, thanks
[15:35] <cjwatson> tseliot: that failed to build because your debian/rules does its hard work in binary-indep but it now needs to be in binary-arch. Shall I just fix it and reupload?
[15:36] <tseliot> cjwatson: weird, it didn't fail here. Please fix it. Thanks again
[15:36] <cjwatson> tseliot: you probably did debuild -b (or no arguments) rather than -B
[15:36] <cjwatson> http://launchpadlibrarian.net/17115976/buildlog_ubuntu-intrepid-amd64.nvidia-common_0.2.2_FAILEDTOBUILD.txt.gz
[15:38] <tseliot> cjwatson: right, I can see the problem now. Sorry
[15:38] <cjwatson> fixed
[15:44] <tseliot> ok
[15:53] <mterry_> Does anyone know if a program can tell when a VT switch happens?  Some event that can be caught?  (I'm looking at UNR launcher bug 237761)
[15:56] <ogra> mterry_, you can put scripts into pm-tools
[15:56] <ogra> that can trigger something for you
[15:57] <ogra> mterry_, /usr/lib/pm-utils/ have a look there
[15:58] <ogra> you can execute any kind of script at resume with that
[15:58] <mterry_> ogra: Hmm, interesting.  But I'm looking for any VT switch, not just resume
[15:58] <ogra> you could check for fgconsole
[15:59] <ogra> grep fgconsole /etc/init.d/usplash
[15:59] <ogra> have a look there
[15:59] <ogra> not sure how to do that from C in a clean way thouh
[16:00] <mterry_> ogra: Also a good tip, but doesn't work from X
[16:00] <cjwatson> there's a VT_WAITACTIVE ioctl you can use to wait for a given console to become active
[16:01] <ogra> mterry_, well, works from X for me ... but needs suid rights
[16:01] <cjwatson> the thread can't do anything else while doing that though
[16:01] <mterry_> ogra: Ah, right you are
[16:01] <ogra> sudo fgconsole properly returns 7 here
[16:01] <mterry_> cjwatson: :-/
[16:02] <ogra> there was a tool to work around the sudo from mjg59 ... not sure we still have it nowadays
[16:02] <mterry_> cjwatson, ogra: Sigh, I was hoping for a nice X event or something
[16:02] <cjwatson> I think there's a consolekit dbus message that gets sent
[16:02] <ogra> mterry_, how about asking the WM about being mapped ... does it really need to be the console ?
[16:03] <ogra> i.e. mapping the wallpaper could just triger your event
[16:03] <ogra> or something like that
[16:05] <bryce> \sh: stay tuned; the code you need isn't in the archive yet
[16:05] <ogra> bryce, i consider touchscreens not working a regression bug and thus not FF critical, do you agree ?
[16:06] <bryce> ogra, bug #?
[16:06] <ogra> none yet, but iÄll file one before FF :)
[16:07] <ogra> esentially it is "no evtouch based touchscreens work after hal-input switch"
[16:07] <bryce> "doesn't work" has such broad meaning...
[16:08] <ogra> bryce, isnt configured at all
[16:08] <mterry_> cjwatson: This consolekit dbus stuff looks promising.  Thanks
[16:08] <ogra> better ?
[16:08] <mterry_> ogra: Not a bad idea either about mapping
[16:08] <ogra> bryce, i have http://paste.ubuntu.com/40402/
[16:08] <bryce> ogra: yes
[16:08] <ogra> bryce, but i have no time to test the matching fdi stuff that needs today
[16:09] <ogra> bryce, works similar to the kbd stuff with sourcing the console defaults
[16:10] <ogra> that way we can ship a common fdi and can have specific setups in /etc/default/evtouch
[16:10] <ogra> if i get around it in time for intrepid i'll fixx the calibration tool to write to /etc/default/evtouch
[16:11] <bryce> I would want to think that evtouch should at least work even if not configured, by using acceptable defaults
[16:12] <ogra> it doesnt
[16:12] <ogra> simply because X likes to pick mouse or evdev for the devices
[16:12] <bryce> anyway, I would think FF wouldn't inhibit posting fixes to a bug like this
[16:12] <ogra> great
[16:14] <cjwatson> dendrobates: who has the baton on getting an updated landscape-client into the archive right now? I gather kirkland is on holiday - has it been handed off to somebody else, or is it going to slip past FF?
[16:14] <cjwatson> dendrobates: I've added the pkgsel question as I understood the specification
[16:15] <dendrobates> cjwatson: I do.  Currently the security team is reviewing it.
[16:16] <dendrobates> cjwatson: if we just uploaded it, it would slip into main without review, due to the empty package alrready there.  We are trying to avoid that.
[16:17] <cjwatson> understood, I just wanted to know whom I needed to sync up with about the installer bits
[16:17] <cjwatson> dendrobates: #ubuntu-installer?
[16:17] <\sh> bryce: you mean screen-resolution-extra and python-xkit?
[16:19] <bryce> yes
[16:20] <\sh> bryce: it's in universe...and I installed it somehow
[16:20] <bryce> also, the screen-resolution-extra has a minor bug I found, that tseliot will be posting a fix for today
[16:20] <bryce> oh great
[16:21] <bryce> it requires a patched gnome-control-center
[16:21] <\sh> ah :)
[16:21] <bryce> http://bryceharrington.org/ubuntu/ScreenRes/
[16:22] <bryce> the bug is that you have to change "time.time()" to "str(time.time()" in one of the python files
[16:22] <\sh> bryce: in the meantime, what's the best way to generate an xorg.conf file, and set the virtual stuff manually?
[16:22] <ogra> bryce, bug #261873
[16:22] <bryce> \sh, https://wiki.ubuntu.com/X/Config
[16:23] <ogra> bryce, hoping my blog entry gets me some more touchscreen info for others
[16:24] <\sh> bryce: thx :)
[16:45] <jdstrand_> why in the world did citadel-server get pulled into my schroot buildd?
[16:46] <cjwatson> Provides: imap-server, pop3-server?
[16:46] <soren> jdstrand_: When building what?
[16:47] <jdstrand_> soren: when dist-upgrading it
[16:47] <soren> Yikes.
[16:47] <soren> I haven't a clue.
[16:47] <jdstrand_> this is my -source
[16:47] <jdstrand_> yeah, me either, no time to figure it out-- it was mostly rhetorical
[16:48] <LaserJock> jdstrand_: do you have devscripts installed in your chroot?
[16:49] <jdstrand_> LaserJock: why yes I do
[16:49] <LaserJock> jdstrand_: that'd most likely be it
[16:50] <LaserJock> Recommends made devscripts pull in quite a bit more than it used to
[16:52] <jdstrand_> LaserJock: I'd say. thanks for the tip
[16:52] <jcristau> jdstrand_: stuff depending on / recommending m-t-a, pulls in citadel-mta
[16:53] <jdstrand_> sure, I just didn't know what that 'stuff' was-- the schroot is pretty barren
[17:04] <BenC> mdz, slangasek, cjwatson: We're discussing the 2.6.27 move and contingency plan on #ubuntu-kernel IRC meeting right now, if interested
[17:07] <dholbach> does anybody know what the problem is in bug 260873?
[17:08] <BenC> cr3: Could I ask you to join #ubuntu-kernel to talk about certification testing with 2.6.27?
[17:09] <cr3> BenC: sure thing
[17:13] <\sh> hmmm..what happend to evolution and its exchange support? it doesn't work in intrepid anymore :)
[17:14] <cjwatson> dholbach: the problem seems to be that ARCH isn't set in Makefile, so ./configure gets --host without an argument, and thus runs out of options during option processing (that's the "can't shift that many" bit)
[17:15] <dholbach> thanks a lot cjwatson
[17:21] <geser> dholbach: see Debian bug 493424
[17:22] <dholbach> geser: great... thanks!
[17:22] <geser> dholbach: testing right now if the new version builds
[17:22] <dholbach> geser: thanks a lot
[17:22]  * dholbach hugs super-geser
[17:23] <jdong> I'm seeing a lot of dscverify can't find debian keyring errors from dpkg -x since Intrepid; is this supposed to happen?
[17:23] <jdong> err, dget -x
[17:23] <jdong> i.e. grabbing a dsc URL off Launchpad; these packages should be signed by Ubuntu
[17:24] <cjwatson> not a valid assumption
[17:24] <cjwatson> for packages we've synced from Debian, they'll be signed by keys in the Debian keyring
[17:24] <cjwatson> we don't re-sign the .dsc
[17:24] <jdong> ok
[17:26] <cjwatson> perhaps somebody should add the Ubuntu keyring to the list of keyrings checked by dscverify, though
[17:26] <jdong> hmm this particular package is an Ubuntu one
[17:26] <cjwatson> (/usr/share/keyrings/ubuntu-archive-keyring.gpg)
[17:26] <jdong> oh is the Ubuntu keyring not recognized by dscverify?
[17:26] <cjwatson> right
[17:26] <geser> dholbach: the new lua-gtk version starts building but fails during the tests :(
[17:26] <cjwatson> however, we don't actually export a list of all keys of all Ubuntu maintainers anywhere
[17:26] <cjwatson> ubuntu-archive-keyring.gpg just has the keys that sign the Release file, not the keys that sign .dscs
[17:27] <cjwatson> so, err, you're screwed right now. best just ignore .dsc verification errors
[17:27] <jdong> cjwatson: thanks for the info :)
[17:28] <dholbach> geser :-/
[17:28] <dholbach> geser: maybe best to just close the sync request for now
[17:29] <geser> perhaps, I've added the last build failure to the sync request
[17:50] <LaserJock> QA Team meeting in 10 min. in #ubuntu-meeting
[18:26] <soren> jdstrand_: Just a friendly suggestion: A dh_ufw might be in order.
[18:27] <jdstrand_> soren: maybe some day, but not today. the dpkg triggers cjwatson suggested worked great
[18:27] <jdstrand_> (after I changed a couple of things)
[18:28] <jdstrand_> soren: but thanks for your suggestion :)
[18:31] <soren> Well, not just for instrumenting postinst, but also to pick them up automagically from the debian directory and put them in the right place.
[18:31] <soren> That way, if we decide "something" needs to be done to ufw rule files, we only need to change it in one place.. But yeah, "some day" is fine. :)
[18:34] <apachelogger> kees: I am not sure if you know already, but the hardening for mysql-dfsg-5.0 broke half of mysql, thus breaking _all_ of KDE
[18:35] <apachelogger> kees: compare binary package list of https://edge.launchpad.net/ubuntu/+source/mysql-dfsg-5.0/5.0.67-0ubuntu1 with https://edge.launchpad.net/ubuntu/+source/mysql-dfsg-5.0/5.0.67-0ubuntu2
[18:39] <cjwatson> apachelogger: it hasn't built yet - StevenK and others were investigating that this morning and we think we have a possible fix
[18:39] <kees> apachelogger: It failed to build only on amd64 on the buildds before hardening, and failed on i386 only after hardening.
[18:40] <kees> apachelogger: both of which I haven't been able to reproduce.
[18:40] <cjwatson> apachelogger: we think in fact the problem was at best only tangentially related to hardening; it's a timing issue in the tests
[18:40] <kees> cjwatson: ah, good to hear
[18:40] <cjwatson> apachelogger: which appears to have been fixed upstream unless I'm much mistaken, so we can pull back that patch
[18:40] <tedg> How do I find what changed in debian policy between two versions?  ie from 3.7.3 to 3.8.0
[18:40] <cjwatson> tedg: upgrading-checklist
[18:40] <saivann> I would need some help with a ubuntu package, can a ubuntu developer help me?
[18:41] <saivann> It's concerning missing files in /usr/lib32 in my x64 package
[18:41] <kees> tedg: install "debian-policy" and look at the upgrading-checklist file in /usr/share/doc/debian-policy/
[18:41] <apachelogger> cjwatson: shouldn't ubuntu3 have fixed the issue then?
[18:42] <apachelogger> we can't testbuild/upload any KDE software right now since it FTBFS
[18:42] <cjwatson> apachelogger: that's quite different from the upstream fix
[18:42] <apachelogger> ok
[18:42] <cjwatson> it is not obvious that simply reverting hardening will reliably fix things
[18:43] <cjwatson> the problems are transient
[18:43] <cjwatson> StevenK: so what happened to that mysql upload?
[18:43] <apachelogger> most awkward kind of problems, due to upcoming FF :)
[18:44] <tedg> Cool thanks kees and cjwatson
[18:44] <cjwatson> kees: perhaps you could grab the patch out of this morning's scrollback and have a look over it
[18:44]  * kees goes digging
[18:44] <cjwatson> http://bugs.mysql.com/bug.php?id=23921 -> http://lists.mysql.com/commits/49751
[18:44] <cjwatson> kees: ^-
[18:53] <kees> zul: were you able to reproduce the failures ever?
[18:53] <kees> zul: I'd like to try reverting your ubuntu3 changes, and applying the upstream patch.
[18:53] <zul> kees: it fails in different places locally never consistenly
[18:53] <zul> kees: cool with me
[18:54] <kees> alrighty, I'll give it a go.
[18:56] <jdstrand_> soren: that's a really good idea. i've put it on my todo list. thanks!
[19:00] <tseliot> slangasek: are you the archive admin today?
[19:07] <soren> jdstrand_: I think dh_installudev will make a good template. If you decide to do it, don't forget to add it to cdbs as well.
[19:08] <tseliot> slangasek: never mind, problem solved
[19:13] <slangasek> tseliot: fwiw, the day assignments are documented on https://wiki.ubuntu.com/ArchiveAdministration
[19:14] <tseliot> slangasek: ok thanks
[19:29] <emgent> hello
[20:18]  * Caesar sniggers at slangasek re #216990
[20:18] <Caesar> Dude, that has just been a disaster from start to finish
[20:18] <slangasek> :(
[20:19] <slangasek> well, lesson learned, that's why pam-config-framework is done for intrepid
[20:19] <Caesar> Cool
[20:19]  * Caesar wishes he had more time to pay attention to Intrepid's dev cycle
[20:20] <Caesar> Unfortunately we're still up to our eyeballs in Hardy deployment shenanigans
[20:22] <slangasek> ah, you should be using update-manager --no-shenanigans
[20:22] <slangasek> I don't know why we didn't make that the default
[20:24] <slangasek> zul: samba 3.2.3-1 in the pipe, includes a security fix (intrepid-only); you're welcome to merge it if you notice it showing up on MoM before I do, or if you opt to do it by hand, or if you know a better way to merge without first waiting for MoM to see it :)
[20:24] <slangasek> (mumble, need an Ubuntu VCS branch for that, mutter)
[20:24] <zul> slangasek: cool on my todo list
[20:24] <Caesar> slangasek: we don't do upgrades, we just reinstall
[20:24] <Caesar> Cuts out a whole slew of problems
[20:25] <slangasek> Caesar: oh, well in /that/ case, I guess --no-shenanigans won't help you ;)
[20:25] <Caesar> :-)
[20:25] <Caesar> Most of our shenanigans are self-inflicted
[20:28] <jdstrand_> zul: are you doing that today? cause I've got the ufw integration bits I need to upload
[20:29] <zul> jdstrand_: as soon as it appears on MoM ill probably do it
[20:30] <jdstrand_> zul: then let me give you a debdiff when I'm done
[20:30] <zul> jdstrand_: no problem
[20:53] <BenC> Anyone know why I can't get any more than 2 workspaces when I have compiz enabled?
[20:53] <BenC> Is there a way to reset (IOW purge) all compiz settings under gconf?
[20:56] <Treenaks> BenC: gconftool-2 's recursive-unset feature should be able to do that
[20:57] <Treenaks> BenC: (you can find the compiz-subtree using gconf-editor)
[21:00] <BenC> Treenaks: thanks
[21:03] <ion_> Which reminds me, i should write a tool that unsets all user’s gconf keys that match the system defaults, so that everything not specifically changed follows any package changes. Unless such a tool already exists, of course.
[21:05] <BenC> ion_: you should just port gconf to use dpkg's conffile handling *snicker*
[21:05] <ion_> :-)
[21:05] <ion_> dpkg’s conffile handling could use some work itself. Perhaps i’ll get around to that some year.
[21:07] <BenC> ion_: there's userspace toolks to handle conffiles better now...can't recall the name of it
[21:08] <ion_> benc: Btw, in case you didn’t notice my recent message, i modified http://heh.fi/patches/grub/01-last-good-boot-update-delay (Sorry for the repetition if you did.)
[21:09] <slangasek> BenC, ion_: ucf
[21:09] <slangasek> which requires you to un-conffile it first, fwiw
[21:09] <BenC> ion_: saw it, thanks
[21:09] <slangasek> (and also badly needs to be integrated with dpkg instead of being a separate package, but it needs other fixes worse than that :)
[21:10] <ion_> benc: I’ll try to remember that you do read all messages. Some people tend to miss messages sent when they were away. :-)
[21:50] <slangasek> jdstrand_: were there other packages that provided auth-client-config profiles in hardy?
[21:50] <slangasek> ah, ldap-auth-config
[21:57] <jdstrand_> slangasek: yeah-- that and ecryptfs were the only two I knew off
[21:57] <jdstrand_> of
[21:58] <mathiaz> kees: did you make any progress with mysql ?
[21:58] <slangasek> jdstrand_: ok.  Do you have an opinion on whether installing libpam-ldap alone should cause auto-configuration of the auth method going forward?  I'm inclined to think it should
[21:59] <mathiaz> kees: apache2 fails to build because of this.
[21:59] <slangasek> jdstrand_: given that, e.g., ldap-auth-config isn't in Debian currently
[22:00] <mathiaz> slangasek: I've uploaded a new version libgems-ruby to intrepid a couple of hours ago. It created a new binary package, which is required by another package (passenger) I'd like to upload to the archive before FF. But as of now passenger doesn't build.
[22:01] <slangasek> mathiaz: so you need a bit of new processing?
[22:01] <mathiaz> slangasek: yop :)
[22:01] <slangasek> ok, looking
[22:02] <kees> mathiaz: yeah, I *finally* got the upstream patch working, and I will upload as soon as it finishes the current testing cycle
[22:02] <kees> mathiaz: apache2> really?
[22:03] <mathiaz> kees: yes - it requires libmysqlclient1.5-off to build
[22:03] <kees> mathiaz: oh!
[22:03] <kees> I thought you meant PIE broke it, which would have be interesting given my testing.  :P
[22:03] <mathiaz> kees: http://launchpadlibrarian.net/17119738/buildlog_ubuntu-intrepid-amd64.apache2_2.2.9-3ubuntu2_FAILEDTOBUILD.txt.gz
[22:03] <kees> mathiaz: yeah, lots of stuff needs mysql -- that's why I've spent all morning working on it.  :)
[22:04] <mathiaz> kees: have you looked at the build failure for i386 ?
[22:04] <mathiaz> kees: I can reproduce it, and it's always the same test that fails (subselect)
[22:04] <kees> mathiaz: that's what this is supposed to fix.  oh!  well then I'll give you this debdiff -- I haven't been able to reproduce it.
[22:05] <mathiaz> kees: are you  refering to http://bugs.mysql.com/bug.php?id=23921 ?
[22:05] <kees> yeah
[22:06] <mathiaz> kees: IIUC this would help in fixing the build failure for amd64
[22:06] <mathiaz> kees: http://launchpadlibrarian.net/17097260/buildlog_ubuntu-intrepid-amd64.mysql-dfsg-5.0_5.0.67-0ubuntu3_FAILEDTOBUILD.txt.gz
[22:06] <mathiaz> kees: http://launchpadlibrarian.net/17049756/buildlog_ubuntu-intrepid-amd64.mysql-dfsg-5.0_5.0.67-0ubuntu1_FAILEDTOBUILD.txt.gz
[22:06] <kees> mathiaz: there wasn't a build failure for amd64.  :)  https://edge.launchpad.net/ubuntu/+source/mysql-dfsg-5.0/5.0.67-0ubuntu2
[22:06] <mathiaz> kees: ^^ these are failures in two differents tests
[22:07] <kees> mathiaz: that's what the mysql upstream fix is supposed to fix -- the random failures being seeing due to timing.
[22:07] <jdstrand_> slangasek: well, I have always had my reservations about that, but that is the point of your work :)
[22:07] <kees> mathiaz: ubuntu1 failed only on amd64.  ubuntu2 only on i386, ubuntu3 only on amd64.
[22:07] <slangasek> jdstrand_: well, it doesn't have to be auto-enabled, and it doesn't have to be in the package that provides the PAM module - it could be in any package that depends on the PAM module
[22:08] <mathiaz> kees: ubuntu3 also failed on i386
[22:08] <slangasek> jdstrand_: but I would think that people shouldn't install the libpam-ldap modle if they don't intend to use it
[22:08] <mathiaz> kees: in the same test as ubuntu2
[22:08] <mathiaz> kees: so it seems we're having two bugs.
[22:08] <jdstrand_> slangasek: that could be said of most any software
[22:08] <slangasek> I say it about most software :)
[22:09] <jdstrand_> heh
[22:09] <kees> mathiaz: hrm, that could be.  I'll get this debdiff to you to test on your failing i386
[22:09] <mia33> lookin 4 cc numbrs
[22:10] <mathiaz> kees: ok. I'll give it a try - I can reproduce the build failure on i386.
[22:10] <kees> mathiaz: can you try building with: http://people.ubuntu.com/~kees/mysql-dfsg-5.0_5.0.67-0ubuntu4.debdiff
[22:10] <slangasek> jdstrand_: the point is, we already have one toggle (package installed or not) that lets people specify their preference; why do we need a second one to declare whether to enable the module, instead of using that one?
[22:10] <jdstrand_> slangasek: if we can safely enable it without breaking a local login, then I suppose that enabling it on install would be ok. that said, ldap-auth-config is really dendrobates thing (though I did quite a bit of packaging work), so I am not sure what his long term intentions are
[22:10] <slangasek> (and, in fact, you can use pam-auth-update to disable the module again for testing)
[22:10] <jdstrand_> slangasek: especially with your latest work
[22:10] <slangasek> oh, ok
[22:10] <slangasek> dendrobates: ping, then :)
[22:12] <dendrobates> slangasek: yes
[22:13] <slangasek> dendrobates: from above, wrt ldap-auth-config, do you see any reason that we shouldn't move the pam part of this down into libpam-ldap directly now that pam-config-framework is in?
[22:14] <dendrobates> slangasek: seems like that would be the correct thing to do.
[22:14] <slangasek> ok, I'll sneak that change in ahead of FF today then :)
[22:18] <liw> I'm having trouble deciding how to treat system-cleaner's state file (contains state of packages the user has indicated should/should not be removed); should I treat it as a config file (i.e., not remove until purge), or as normal application data (remove when package is removed, not wait until purge)
[22:18] <liw> opinions?
[22:19] <pwnguin> well, given the goal is to reduce disk usage
[22:19] <pwnguin> make that easy as possible ;)
[22:20] <pwnguin> then again, I dont think I'd use the tool
[22:20] <pwnguin> you should seek opinions from your users!
[22:25] <slangasek> mathiaz: libgems-ruby accepted
[22:27] <mathiaz> slangasek: thank you ! :)
[22:31] <dendrobates> slangasek: so landscape-client is currently being reviewed by kees/jdstrand.   What other review do you want?  There is a empty package in main, and I don't want this to slip in unreviewed.
[22:31] <slangasek> dendrobates: well, MIR is not really my department...
[22:31] <slangasek> dendrobates: if it's gotten the security review, then that's the main thing for the MIR team, I think
[22:33] <ogra> slangasek, but post MIR is your dept. right ?
[22:33]  * ogra is about to add sshfs to ltsp-client and this to the cd
[22:33] <slangasek> uhm, post-FF feature changes are my department
[22:33] <ogra> *thus
[22:34] <slangasek> to which CD?
[22:34] <ogra> alternate
[22:34] <ogra> its about 300k or so
[22:34] <slangasek> and no new deps?
[22:34] <slangasek> I guess we can spare 300K :)
[22:34] <ogra> not afaik
[22:34] <ogra> good :)
[22:34] <ogra> MIR was approved but i'm late as usual with ltsp for FF
[22:35] <dendrobates> slangasek: I want you to be aware that I am holing off on uploading this until the security review, so it doesn't slip into main.  This will mean a ffe.
[22:35] <slangasek> dendrobates: sure; jdstrand_ talked to me about that yesterday already
[22:36] <ogra> https://bugs.launchpad.net/ubuntu/+source/sshfs-fuse/+bug/254651 well, ubuntu-archive is you :)
[22:36]  * ogra just noticed it was moved on
[22:39] <ogra> YIPPIEEEE !!!!
[22:39]  * ogra dnces
[22:39] <ogra> *dances too :D
[22:39] <ogra> finally ....
[22:40] <ogra> took me the whole day to finally get hal-input working in ltsp
[22:50] <persia> I'm well behind on backscroll, but if anyone is still trying to sort out the perl 5.10 and threads issue, there's a package on the RCbugs page that appears to be perl and threads and has a new upstream in Debian.
[22:51] <kees> mathiaz: any love for the mysql build yet?
[22:52] <mathiaz> kees: the test are currently running
[22:53] <mathiaz> kees: I should have a answer in a few minutes
[22:53]  * kees really hopes this works.
[22:57] <mathiaz> kees: nope - the subselect test failed
[22:58] <mathiaz> kees: I've search mysql bug tracker for reference about pie and subselect on i386, but nothing came out
[22:58] <kees> mathiaz: well... that sucks.  let me see if I can reproduce here again...
[23:05] <kees> mathiaz: are you running an intrepid kernel?
[23:06] <kees> mathiaz: if so, are you seeing any segvs in dmesg for the server?
[23:06] <mathiaz> kees: nope - I'm running sbuild on a hardy server
[23:06] <kees> amd64?
[23:06] <mathiaz> kees: yes
[23:06] <kees> it should show i386 segvs too in hardy... nothing in dmesg?  (apport would ignore it since the test server isn't packaged)
[23:06] <mathiaz> kees: that's the command I use to build mysql:  sbuild -A -d intrepid-i386 mysql-dfsg-5.0_5.0.67-0ubuntu4.dsc
[23:07]  * kees nods
[23:07] <mathiaz> kees: nothing
[23:07] <kees> well, hopefully I can reproduce, but that this point, I'm leaning towards ripping out PIE and dealing with it in intrepid+1.
[23:08] <kees> the "Lost connection to MySQL server" is ominous.
[23:15] <jdstrand_> zul: debdiff in bug #261544 is ready for you
[23:17] <jdstrand_> slangasek: I don't know if you or zul is preparing the samba upload, but if you do, can you incorporate ^^
[23:18] <slangasek> jdstrand_: zul said the magic words "it's on my TODO list", so I was going to cede it to him :)
[23:18] <jdstrand_> heh, fair enough :)
[23:22] <ogra> slangasek, may i point you kindly to bug #262036 ?
[23:53] <kees> [244640.179532] mysqld[11269]: segfault at 57aafec8 ip 55f220e0 sp 57aafe9c error 6 in libc-2.8.90.so[55ee4000+13d000]
[23:54] <ion_> Those hex numbers translate to ASCII “Switch to Postfix”. ;-)
[23:54]  * ion_ slaps himself
[23:54] <ion_> Postgres, duh.
[23:54] <kees> hehe.
[23:55] <kees> actually, the "6" translates to: No page found (PF_PROT==0) | Write (PF_WRITE==1) | Userspace (PF_USER==1) | No reserved bit (PF_RSVD==0) | Regular memory access (PF_INSTR==0)
[23:56] <tormod> slangasek: can you please take a look at bug 192772? either sync or merge is fine. I have given up on rtg.
[23:57] <ogra> tormod, hey
[23:58] <tormod> ogra: hey
[23:58] <ogra> tormod, not sure where tedg is at atm, i asked him for a xscreensaver package before feature freeze today but seems nothing happened, are your changes in debian ?
[23:58] <ogra> (i assumed so)
[23:58] <tormod> ogra: tedg promised me to do a 5.07, but he left IRC and I see nothing
[23:58] <ogra> tormod, right, same here :/
[23:59] <ogra> tormod, any intrest in becoming ubuntu developer ? :)
[23:59] <tormod> ogra, basically I have prepared 5.07-1, but it's not released.
[23:59] <tormod> ogra: interest yes, but I don't know about time :)
[23:59]  * ogra would really like to see the screensaver situation in one hand in both distros ...