[00:30] <Persica> Sorry, new here.  What's the deal with automount on nfs partitions?  I was told it was patched to not work on an NFS partition, but it seems to try to start just fine on my NFS root.
[00:30] <Persica> However, it won't mount anything after that...
[00:42] <Persica> ah well, back to hacking at it.  I guess I have to dive deeper.
[00:56] <Persica> oooh, is mkdir_path trying to do a stat on the directory and not getting a nice response from automount?
[01:11] <slangasek> superm1|away: so it seems we have an issue with the amd64 livefs buildd hanging whenever I try to do anything with it, I guess DVD builds are going to have to wait until someone is available to unblock this
[01:11] <slangasek> (and hopefully to diagnose it)
[01:21] <jdong> lpget vlc/intrepid
[01:25] <hardwire> whats lpget?
[01:29] <pwnguin> probably a personal script to grab changes/packages/source/bugs from launchpad
[01:32] <jdong> yeah, it's a python script to scrape LP for dsc links
[01:32] <jdong>  3461 files changed, 1495678 insertions(+), 1152722 deletions(-)
[01:32] <jdong> boy this'll be fun.
[01:35] <pwnguin> surely most of that is moved files
[01:35] <wgrant> jdong: Have you asked Debian if they've had a look?
[01:36] <wgrant> We're basically in sync with them now.
[01:36]  * RAOF blinks at 1.4x10^6 LOC changed
[01:38] <jdong> wgrant: no, I haven't.... that would be a good idea though; right now I'm just trying to gain an understanding on the extent of the work required
[01:42] <wgrant> jdong: Approximately infinite.
[01:43] <wgrant> Or as close as a new minor version can get to infinite.
[01:43] <jdong> wgrant: yeah it looks a bit painful indeed
[01:43] <jdong> nonetheless, it's this or math homework.
[01:47] <jdong> well, it seems to be getting through configure....
[01:47] <jdong> of course I need to go back and s/drop/rewrite/ these patches.
[03:22] <ender_> I have a data alignment question. I noticed that addresses returned from malloc() are always aligned (end in 0 hex) to the processors native data type?
[03:22] <ender_> Is this something I can rely upon, i.e. it will always be the case?
[03:26] <ender_> Nevermind... found the answer: "The address of a block returned by malloc or realloc in the GNU system is always a multiple of eight (or sixteen on 64-bit systems)." Very cool.
[04:08] <j-b> hello
[04:09] <j-b> what should I tag on launchpad when a bug is fixed upstream ?
[04:19] <persia> j-b: You'd want to ask that sort of question in #ubuntu-bugs in the future.  TO indicate fixed-upstream, add a bug task for the upstream project, and link it to the upstream bug that is fixed.
[04:20] <j-b> persia: ok, thanks...
[04:20] <j-b> persia: not going ot be fixed before 10 month then :D
[04:22] <persia> j-b: Hrm?  If it's fixed upstream, and it's marked as described, it appears in the list of bugs fixed upstream.  There are a few developers that regularly review these lists, and try to pull the fixes.
[04:25] <yao_ziyuan1> i just hope intrepid can have a working netboot installer
[04:25] <yao_ziyuan1> because my cd burner is not currently supported by linux kernel
[04:34] <persia> yao_ziyuan1: Have you tried a netboot from one of the Alpha release candidates?
[04:34] <yao_ziyuan1> no
[04:34] <yao_ziyuan1> but i have nightmares with netboot in hardy
[04:35] <yao_ziyuan1> as it said, netboot in hardy didn't work
[04:35] <persia> yao_ziyuan1: Best path to make sure it works in intrepid is to test the netboot installer from the release candidates, and report bugs.
[04:36] <yao_ziyuan1> persia: are you from iran...
[04:36] <persia> yao_ziyuan1: Nope.
[04:38] <yao_ziyuan1> ok. i'll make a snapshot of my windows virtual machine in vbox,
[04:38] <yao_ziyuan1> and then try the netboot files of intrepid in this windows vm
[04:39] <yao_ziyuan1> using grub4dos
[04:39] <yao_ziyuan1> is that ok?
[04:39] <persia> yao_ziyuan1: Not being someone who has ever tried a netboot install from Windows, I'm not sure, but there's usually no problems with testing in virtual machines.
[04:40] <yao_ziyuan1> i'm installing intrepid alpha 5 in vbox as a new vm
[04:40] <yao_ziyuan1> using the alternate installer iso
[04:41] <yao_ziyuan1> maybe after installation i can take a snapshot and try netboot in it
[05:56] <dholbach> good morning
[05:57] <TheMuso> hey dholbach
[05:57] <dholbach> hiya TheMuso :)
[05:57] <ion_> Hi
[05:57] <dholbach> hi ion_!
[05:57] <ion_> What’s up?
[05:58] <dholbach> I'm slowly waking up :)
[05:58] <dholbach> how are you doing?
[06:03] <ion_> I wrote a small program that tracks computer activity; by analyzing its log i can automate a sleep diary. :-)
[06:03] <ion_> http://heh.fi/tmp/idlemeter
[06:41] <Q-FUNK> howdy!  could someone please look into bug #194140 ?  it's been sitting there unattended for ages.
[07:09] <persia> Anyone read Russian well enough to understand whether bug #270773 is valid?
[07:09] <persia> -ECHANNEL
[07:11] <Treenaks> persia: google translate translates it pretty ok
[07:55]  * NCommander is awake!
[08:01] <NCommander> TheMuso: Good mroning
[08:01] <TheMuso> Hey NCommander.
[08:02] <NCommander> TheMuso: can I steal you for an ack on a package (its got a dependency change, I'm already got one ack from -release, but a second one would be nice so we could finally get this uploaded)
[08:03] <TheMuso> NCommander: bug number?
[08:03] <NCommander> TheMuso: https://bugs.edge.launchpad.net/ubuntu/+source/gnucash/+bug/270200
[08:04] <TheMuso> looking.
[08:06] <siretart> TheMuso: since when do we require freeze exception for patches that change dependencies like that bug?
[08:07] <TheMuso> siretart: thats a good question.
[08:07] <siretart> according to https://wiki.ubuntu.com/FreezeExceptionProcess#Exceptions%20for%20Universe/Multiverse, we don't
[08:08] <persia> I don't think it needs a freeze exception, but that specific bug probably needs several pairs of eyes, as there are historical reasons why it still uses 0.4
[08:08] <TheMuso> Interesting that ScottK didn't just mark it confirmed/upload it.
[08:08] <NCommander> TheMuso: he wanted a second ack
[08:09] <TheMuso> Oh ok.
[08:09] <NCommander> We had three testers from the gnucash team test it in addition to myself. The upgrade also resolves quite a few bugs
[08:09] <NCommander> (related to PDF and fonts)
[08:10] <TheMuso> NCommander: SO it works?
[08:10] <Q-FUNK> what are the changed between kernel packages for 2.6.24 and 2.6.27 that are needed to get the vcons back?  evidently, yet more changes to the way fbdev modules are packaged, but which ones?  is the strategy for this documented anywhere?
[08:11] <NCommander> Yeah. It gained a new bug though because it uses the GTK print system now vs. its old one
[08:11] <slangasek> while you're at it, can someone fix the annoying glitch that I can only select a reconciliation date in gnucash using the calendar widget, not by entering a date string?
[08:11] <NCommander> TheMuso: That bug however affects all GTK applications
[08:11] <NCommander> TheMuso: and has already been reported on LP and upstream
[08:12] <TheMuso> NCommander: Ok.
[08:13] <TheMuso> NCommander: Ok ACK.
[08:13] <TheMuso> NCommander: in the bug also.
[08:13] <NCommander> TheMuso: any chance I can convience you to upload as well ;-)?
[08:14] <TheMuso> NCommander: Yep I'll do that as well.
[08:14] <NCommander> \o/
[08:16]  * NCommander starts attacking the libtool fixs
[08:16] <StevenK> Hmph
[08:16] <StevenK> You know, I was looking at libtdl3-dev NBS
[08:19] <NCommander> I was working on ksimus
[08:19] <NCommander> If you want I'll look at another one
[08:20] <StevenK> I have a whole bunch that build
[08:20] <StevenK> I can upload them
[08:20] <NCommander> Which ones are left
[08:20] <TheMuso> Don't remind me about libltdl/libtool. :S
[08:21] <NCommander> At least powerpc doesn't usually break as often as other libtool architectures TheMuso
[08:21] <TheMuso> NCommander: Possibly, I haven't dug that deeply.
[08:21] <NCommander> StevenK: ?
[08:21] <StevenK> NCommander: Hold on
[08:22] <StevenK> I looked at the list like a week ago.
[08:22] <StevenK> cacao freeradius gnash guile-gnome-platform ion2 ksimus myodbc openc++ openct pacemaker php-mcrypt snort
[08:22] <StevenK> That's what I currently have as not building
[08:22]  * NCommander grabs cacao
[08:22] <persia> I thought now that we had ion3 we didn't need ion2 anymore.
[08:23] <StevenK> I have 15 here that build successfully, but I have not uploaded them, or looked at which ones need uploading.
[08:23] <StevenK> Yes, ion2 has been removed
[08:23] <slangasek> ion2 isn't there anymore, yes
[08:24] <NCommander> Ack
[08:24] <ion_> People who didn’t feel like using ion3 after it became non-free probably switched to some other tiling window manager than ion2.
[08:24] <NCommander> cacao is just evil
[08:24] <persia> cacao is essential.  Please fix it :)
[08:24] <StevenK> freeradius looks odd, since it builds, and then exits 1 from debian/rules because openssl has been pulled in somehow
[08:25] <NCommander> StevenK: its a faulty test
[08:25] <NCommander> StevenK: I was helping to debug it. The test it detecting a non-existant openssl linkage
[08:25] <StevenK> Hah
[08:25] <NCommander> I thought a patch was already uploaded to disable said test
[08:25] <NCommander> Guess now
[08:25] <NCommander> *not
[08:25] <NCommander> Well, cacao fixed in experimental, but its a new upstream version
[08:25] <NCommander> *twich*
[08:29]  * NCommander looks at backporting
[08:29]  * StevenK looks at which packages built
[08:30] <TheMuso> NCommander: uploaded.
[08:31] <NCommander> \o/
[08:31] <StevenK> Oh, grumble.
[08:31] <NCommander> ?
[08:31] <StevenK> I suspect someone has already uploaded them
[08:33]  * StevenK sighs
[08:39] <NCommander> Is openjdk available?
[08:47] <NCommander> StevenK: the cacao build is really quite broken
[08:48] <NCommander> StevenK: anyway, the problem with freeradius is one of its build-deps on Ubuntu is built with SSL while it isn't on Debian, and thus it gets linked against openssl even though its using none of ssl's symbols
[08:49] <persia> NCommander: OpenJDK hasn't been ported to powerpc at last report.
[08:49] <StevenK> NCommander: I figured that
[08:49] <NCommander> O_o;
[08:49] <NCommander> We have a port of openjdk to m68k, but not powerpc?
[08:49] <persia> (upstream would welcome patches, but won't do it without someone else getting involved)
[08:49] <NCommander> WTF?
[08:51] <NCommander> persia: think you can answer a license question?
[08:51] <NCommander> libsnmp depends on openssl
[08:51] <NCommander> libsnmp is new-style BSD license so it can be safely linked
[08:52] <NCommander> Can I then take that libsnmp w/ SSL and link it against freeradius?
[08:52] <liw> what's the license of freeradius?
[08:52] <geser> NCommander: net-snmp is also on Debian linked with libssl
[08:53] <persia> NCommander: I'm willing to offer an opinion, but no warranty that said opinion is meaningful in any given jurisdiction, no assurance that said opinion represents informed counsel, nor does my granting of an opinion imply that it matches the opinions of the archive-administrators.
[08:53] <geser> but in Debian "-lsnmp" doesn't get translated to "/usr/lib/libsnmp.so -lcrypto" during build but only to "/usr/lib/libsnmp.so"
[08:53] <geser> during the build by libtool
[08:54] <NCommander> On Ubuntu the pc file pulls into openssl
[08:55] <slangasek> hit the .pc file with a hammer?
[08:55] <persia> NCommander: My opinion is that you can't link it.
[08:55]  * NCommander likes slangasek's suggest
[08:55] <geser> liw: the main binary seems to be GPLv2 and the plugins LGPL
[08:55] <NCommander> persia: which is mine too, but I wanted a second opinion
[08:57]  * NCommander notes openssl is evil
[08:57] <NCommander> evil like pie
[08:58] <NCommander> or cake :-P
[08:58] <liw> openssl is lethal to diabetics?
[08:58] <NCommander> I was thinking more that you can never just take one slice of pie
[08:58] <NCommander> It just sits there taunting you
[08:58] <NCommander> Daring you to eat it
[08:59] <NCommander> And when you do. BAMN, 20 pounds of fat you didn't have before
[09:00] <yao_ziyuan1> seems the intrepid netboot installer can't respond well to network problems
[09:00] <yao_ziyuan1> in the stage "Loading additonal components",
[09:00] <yao_ziyuan1> i paused the vm and saved its state and closed virtualbox
[09:00] <yao_ziyuan1> and then restarted the host machine (ubuntu 8.04 + kde 4.1.1)
[09:01] <yao_ziyuan1> and then restored the vm
[09:01] <yao_ziyuan1> now the netboot installer stalls at "Retriving crypto-modules-2.6.27-generic-di" (4%) forever
[09:02] <yao_ziyuan1> i have to restart the netboot process
[09:03] <yao_ziyuan1> i remember gutsy's netboot installer can give me a Retry button when it fails to download a package
[09:03]  * NCommander sees if libsnmp will link against gnutls
[09:04] <yao_ziyuan1> so i wonder why such good behaviors are not inherited by hardy's and intrepid's netboot
[09:05] <yao_ziyuan1> hoho,
[09:05] <yao_ziyuan1> it does provide a Retry dialog!
[09:05] <yao_ziyuan1> good.
[09:05] <yao_ziyuan1> i intentionally disabled the network interface of this virtual machine.
[09:07] <NCommander> I think this is a serious bug, due to libsnmp linking against openssl
[09:14] <NCommander> slangasek & StevenK: So it seems freeradius has been indirectly linking to openssl for years ....
[09:15] <slangasek> not in Debian; I don't know about in Ubuntu
[09:15] <NCommander> slangasek: libsnmp in debian is linked to libssl
[09:15] <NCommander> And the changelog puts that change in '05 if not earlier
[09:17] <slangasek> hmm.
[09:17] <slangasek> I don't think the freeradius maintainers were aware of that, then
[09:17] <NCommander> YEah
[09:17] <NCommander> Since they have an explicate check for freeradius to see if it was linked against openssl
[09:18] <NCommander> A change in the Ubuntu snmp caused freeradius to pull in the -lcrypto linking flag
[09:18] <NCommander> (for some reason -lcrypto wasn't in the pc before for libsnmp)
[09:18]  * slangasek mutters about Libraries.private
[09:19] <slangasek> or Libs.private, or whatever
[09:19] <NCommander> yeah
[09:19] <AnAnt> Hello, is there anything I should add to this bug 269855
[09:19] <NCommander> And since libssl didn't get pulled in since it wasn't directly linked
[09:19] <NCommander> No lintian warning bells went off
[09:20] <AnAnt> is Ian Jackson here ?
[09:20] <wgrant> Doesn't look like it.
[09:21] <persia> AnAnt: Frequently, unless there is a specific reason you need a specific person, you'll get a better response asking the question generally.
[09:21] <AnAnt> persia: I did
[09:26] <geser> NCommander: where did you find a .pc file for libsnmp?
[09:27] <NCommander> geser: during my debugging
[09:27] <NCommander> Although I don't seem to having it now in pkgconfig
[09:27]  * NCommander grumbles
[09:31] <NCommander> WOOOOO
[09:31] <NCommander> I got a two-for-one license issue today
[09:31] <NCommander> net-snmp is missing the openssl notifications in their copyright
[09:36] <slangasek> huh? why should they have them?
[09:37] <doggymenz> It is very important that Adobe Flash gets updated from -beta to -rc, and that VLC gets updated from 0.8.6 to 0.9.2 before the release of 8.10 Ibex
[09:37] <NCommander> slangasek: clause six of the openssl license
[09:37] <doggymenz> the -beta version of Flash is horrible slow in fullscreen mode, and has crappy performance, this is fixed in the -rc
[09:37] <NCommander> http://paste.ubuntu.com/47375/ - how's that look? (I just want to make sure I'm not misexplaining the issue)
[09:37] <slangasek> NCommander: net-snmp is not a redistribution of OpenSSL.
[09:38] <NCommander> It includes bits of openssl via linking
[09:38] <NCommander> To me thats redistribution "of any form whatsoever"
[09:38] <slangasek> not copyrightable bits.
[09:38]  * NCommander just caught a nice typo in his bug email
[09:39] <NCommander> Hrm
[09:39] <wgrant> doggymenz: What is so critical about breaching FeatureFreeze for vlc 0.9.2?
[09:39] <NCommander> In that case, then freeradius can safely indirectly link to openssl, although net-snmp still needs the adversing notice as per section 3: http://openssl.org/source/license.html
[09:39] <doggymenz> wgrant, the 0.9 branch is the long awaited update of VLC 0.8 branch, it brings much new stuff
[09:40] <wgrant> doggymenz: It is well past feature freeze.
[09:40] <slangasek> no, freeradius can't safely indirectly link to openssl; the GPL requires that when you ship a binary, the source to "all modules it contains" be made available under the terms of the GPL
[09:40] <doggymenz> :(
[09:40] <doggymenz> wgrant, ok, but flash can get new version?
[09:41] <wgrant> "It has some new buggy features" doesn't sound like a very compelling reason to update to a new version.
[09:41] <NCommander> slangasek: ok, let me fix my reasoning then
[09:41] <wgrant> Flash I'm not sure about.
[09:41]  * NCommander bows down to slangasek's expert knowledge of licenses
[09:41] <slangasek> and this has long been intepreted in Debian as meaning that GPL apps can't dynamically link against GPL-incompatible libs within the archive
[09:41] <slangasek> s/licenses/Debian traditions/...
[09:42] <NCommander> I knew they couldn't do that directly, but I didn't know for sure via an indirect linkage
[09:43]  * NCommander fixs his email up
[09:44] <asac> anyone sees a reason why devscripts shouldnt have ca-certificates in Recommends?
[09:45] <slangasek> why /should/ it have it?
[09:45] <slangasek> what does devscripts have in it that cares about SSL?
[09:45] <asac> slangasek: dget from launchpad
[09:45] <NCommander> slangasek: http://paste.ubuntu.com/47378/ - would you mind looking this over to make sure I haven't missed anything important?
[09:46] <asac> slangasek: bug 247157
[09:46] <slangasek> NCommander: sorry, I don't really have time to right now
[09:46] <NCommander> ah, it
[09:46] <NCommander> s/it/ok/g
[09:46] <slangasek> asac: that's filed against ubuntu-dev-tools though, not devscripts?
[09:46] <asac> slangasek: no its against both
[09:46] <asac> slangasek: ubuntu-dev-tools task is fixed. devscript has still a pending patch
[09:47] <asac> which i planned to sponsor ;)
[09:47] <slangasek> asac: ah.  well, I would argue that relying on LP SSL to verify .dsc files is the wrong trust model
[09:48] <asac> slangasek: i think the idea is that you can easily dget from launchpad links
[09:48] <slangasek> sure you can
[09:48] <NCommander> dsc files already have a GPG signature, why can't you simply verify it?
[09:48] <asac> (not to verify, but just to make it the https:// links work)
[09:48] <slangasek> er, there are ways to make https:// work without pulling in the certificates, though?
[09:49] <NCommander> dget uses curl or wget internally, wget has --ignore-certificates (or something like that)
[09:49] <NCommander> Curl probably has an equivelent switch
[09:49] <asac> slangasek: yes there are. but why not make it work out of the box ;)
[09:49] <slangasek> I think we should make it work out of the box, without worrying about certificate validation...
[09:49] <asac> but i am not really personal affected to it. just wanted to ask if someone has real concerns
[09:50] <asac> slangasek: ok. I will ask him to make dget ignore certificate problems instead then (dumping a warning maybe?)
[09:50] <asac> is that an accurate reflection of what your idea is?`
[09:50] <slangasek> practically speaking, this isn't an issue for desktop users; and in my devel chroots I disable recommends; so I guess I don't have a strong argument why ca-certificates /shouldn't/ be depended on, it just seems to me that it's not the simplest solution
[09:51] <slangasek> asac: I think that's accurate, yes
[09:52] <asac> ok thanks for your input
[09:56] <asac> slangasek: you submitted a patch to NM right? what was the bug id?
[09:57] <slangasek> asac: 269010
[09:59] <asac> ok
[10:00] <asac> slangasek: good catch
[10:01] <slangasek> asac: took me much longer than it should've to actually catch it ;P
[10:02]  * slangasek wanders bed-wards
[10:02] <asac> yeah. but you usually wouldnt expect such a glitch ;)
[10:03] <asac> slangasek: isnt freeze for a-6 today?
[10:03] <slangasek> asac: yes
[10:03] <asac> when you wake up?
[10:03] <slangasek> yes
[10:03] <asac> or now?
[10:03] <slangasek> :)
[10:03] <asac> good ;)
[10:03] <slangasek> I milestoned this bug anyway ;)
[10:04] <mvo> +7h to upload crack^Wimportant fixes
[10:04] <slangasek> mvo: didn't you hear?  The Feature Freeze now also includes a Crack Freeze
[10:04] <mvo> hrm, that is a real blow :)
[10:13] <asac> slangasek: ok committed and pushed upstreawm
[10:13] <slangasek> cheers
[10:25] <asac> ogra: meister :)
[10:25] <ogra> asac, dude !
[10:25] <ogra> whats up ?
[10:26] <asac> göttingen ;)
[10:27] <ogra> asac, kommste ?
[10:28] <asac> ogra: thats the question :) -> priv msg
[10:30] <ogra> yep
[10:32] <davmor2> query on todays iso image.  Why is there a listing for other that contains a file call compiz that seems to do nothing?
[10:32] <davmor2> in the main Applications menu sorry ^
[10:36] <mvo> davmor2: that sounds like a bug, let me fix it
[10:41] <davmor2> mvo: okay cool do you want me to report so you can cross it off once fixed?
[10:43] <mvo> davmor2: its already in bzr, no need to file a bug (unless you want to file one :)
[10:43] <mvo> (in bzr now)
[10:44] <asac> cjwatson: working on sponsoring bugs i looked at bug 258036 ... the author already updated his patch. do you want to finish this or shall I?
[10:45] <davmor2> mvo: Okay I'll leave it and write one if it's still in place for alpha 6
[10:48] <cjwatson> asac: go ahead
[10:48] <NCommander> asac: were you working on fixing the EULA bug in abrowser?
[10:54] <asac> NCommander: i dropped a comment in bug 269795
[10:54] <NCommander> asac: I found a patch to fix it
[10:55]  * NCommander opens a bug to request firefox moved to restricted as not free software as you can't modify it freely unless you disable branding
[10:55] <NCommander> Regardless of the EULA, as long as that trademark policy is in place, its not free
[10:58] <NCommander> asac: er, how is it belong to ubufox, an extension can't prevent a tab from opening
[11:00] <cjwatson> NCommander: may be worth noting that even the DFSG permits licences that require derived works to change their name
[11:00] <cjwatson> although that apparently wasn't carried over to the Ubuntu licensing policy, but I think that was an oversight
[11:00] <NCommander> cjwatson: yes, if the version in Debian changes its name
[11:00] <cjwatson> NCommander: the DFSG says no such thing
[11:00] <cjwatson> http://www.debian.org/social_contract point 4
[11:01] <NCommander> Debian packages are built with the orig and a diff, which by definition is a modification to the original package, and in my eyes, a derived
[11:01] <cjwatson> I think that is at best unclear. For a long time Apache technically required a name change on modification ...
[11:02] <NCommander> The reason I state this point is because under section 4, it says that the license must allow modification of the source at build time.
[11:02] <cjwatson> which firefox does
[11:03] <cjwatson> look, I don't particularly like the firefox situation, but you're making an overbroad claim
[11:03] <cjwatson> the obvious counterexample to the claim that every part of the upstream source must be modifiable is that you are not typically permitted to modify copyright notices
[11:03] <NCommander> The way I see it is that when you build a package, you apply a diff to the original source, and then get debs from it. That from me counts as a modification.
[11:04] <NCommander> and if its a modification, then the rules applying to "may require a derived work to carry a different name or version number" appears to apply
[11:04] <cjwatson> that depends on how extensive the requirement to carry a different name is
[11:04] <cjwatson> if it's conditional, and if that condition is extended equally to Debian (or Ubuntu) and everyone else, then it doesn't violate DFSG#8
[11:05] <NCommander> Last time I checked, firefox allows no modifications without prior consent
[11:05] <NCommander> clause eight is why Firefox was eventually removed from Debian in place of Iceweasel
[11:06] <NCommander> They were granted a license for the trademark with the modifications to build a deb, but it was rejected due to 8 since it was Debian specific
[11:07] <cjwatson> of course, the Ubuntu licensing policy explicitly states that non-code elements will be considered case-by-case
[11:08] <cjwatson> when founding Ubuntu we explicitly didn't want it to descend into interminable debian-legal arguments
[11:08] <cjwatson> (and non-code elements were a big issue at the time)
[11:08] <NCommander> Normally I'd agree, but the trademark prevents code modifications. You can't take a firefox source package, modify it, and redistribute it legally
[11:09] <NCommander> You need to disable branding in the rules, and then rename the source package to remove the trademark
[11:09] <cjwatson> that is an issue and it's something I and others are trying to sort out. That doesn't mean moving it to restricted RIGHT NOW is going to help
[11:10]  * NCommander backs down from doing anything
[11:10] <NCommander> I trust your judgement on such matters :-)
[11:10] <asac> NCommander: the source package name shouldnt be an issue
[11:10] <asac> NCommander: and afaik not even the package name and the binary name can be trademarked
[11:11] <asac> NCommander: its really just about the marks presented in the UI
[11:11] <NCommander> asac: unless I'm mistaken, the trademark extends to the name "Firefox"
[11:12] <asac> NCommander: the package name is a functional element which shouldnt be covered by trademarks. but IANAL.
[11:12] <NCommander> Neither am I
[11:12] <NCommander> So I shall put that aside
[11:13]  * NCommander hates to sound like a zealot, but this FF **** just pisses me off
[11:13] <asac> the package description might be an issue if it wrongly claims that this _is_ firefox. we are certainly open for discussion how to improve this. and actually i am working with gnewsense folks to make the package more useful even for distros that cant ship the "firefox" source
[11:14] <NCommander> asac: anything I can do to help?
[11:15] <ogra> how are we communicating that to our derivatives ?
[11:15] <asac> NCommander: you can join #ubuntu-mozillateam and contribute ;) i dont like the idea to do radical packaging changes for intrepid, but we could certainly experiment with the packaging of our 3.1 branch
[11:15] <asac> ogra: is that a question in my direction?
[11:16] <NCommander> Well, I'd like to at least make icecat available via a PPA
[11:16] <asac> ogra: gnewsense "firefox" dev is in my channel and we are actively working on it
[11:16] <ogra> asac, well, a general one ... if a derivative team comes and uses the ubuntu CD as a base ...
[11:16] <NCommander> Just because it disables non-free plugin sensing, which while for me is not important, people who want to run all free systems would apperiate it
[11:16] <asac> NCommander: that can be done in a different fashion
[11:16] <ogra> asac, they should know what they can do and what they cant without violating laws ...
[11:17] <NCommander> asac: ?
[11:17] <asac> NCommander: we have the plugin finder wizard. derivitaves can just have their own database and dont show any non-free plugin there
[11:17] <asac> NCommander: it is even trivial to provide that as a service from our database
[11:17] <asac> because we have the "section" in the database
[11:18] <asac> so ubufox derivitaves can just pass a special plugin finder url parameter and then dont get results for non-free plugins
[11:18] <NCommander> Ah, I didn't know
[11:18] <asac> i am working on that topic the gnewsense too ;)
[11:18]  * NCommander has never looked deep within the depths of Gecko :-)
[11:18] <asac> NCommander: the other feature is available in a icecat extension
[11:18] <NCommander> I thought icecat worked by modifiying the firefox source
[11:18] <asac> which someone should embrace in our firefox extension packaging community
[11:19] <NCommander> If its an extension, I think I'll be shocked silly
[11:19] <asac> NCommander: thats how icecat works. but that isnt the smartest solution
[11:19] <asac> NCommander: not all is in an extension. i presented the idea of a "free" plugin finder database to icecat developers long time ago
[11:19] <asac> and also suggested that they should try to put as much as possible in an extension instead of forking upstream code
[11:20] <asac> they came back to me and said they had an extension that at least covers partso f the use cases
[11:20] <NCommander> I've only written IE extensions, I know nothing of NSAPI :-)
[11:20] <NCommander> (expect that its probably less braindamaging)
[11:20] <asac> NCommander: NSAPI != extensions
[11:20] <NCommander> oh right
[11:20] <asac> NSPAI == content plugins (aka flash, totem)
[11:20] <NCommander> Sorry, I'm still the old school plugins == extensions
[11:21] <asac> NCommander: extensions is more like writing web-pages ... just using xul
[11:21]  * NCommander has bad experiences with xul
[11:21] <asac> of course tricky things that want to tweak stuff in the inner guts are not that simple
[11:21] <NCommander> Then again, I ran xulrunner on m68k ....
[11:21] <asac> NCommander: which is a stupid idea ... without questions asked ;)
[11:22] <NCommander> I had to fix a portability bug ;-)
[11:22] <NCommander> asac: I take it you perfer gecko over webkit then?
[11:22] <asac> NCommander: anyway. packaging extensions is simple. join #ubuntu-mozillateam and help out ;) ... i can connect you to the icecat developer and tell him that you would be happy to help getting his icecat extension into the ubuntu archvie ;)
[11:23] <NCommander> I .... err ... what?
[11:23] <asac> NCommander: i have no preference preference for the rendering engine part. only thing i can say that webkit never made an official/supported release
[11:23]  * NCommander finds webkit somewhat unstable
[11:24] <NCommander> On Linux
[11:24] <asac> NCommander: its not even that it might be unstable. it doesnt have any track record of how they do stable support (read: security)
[11:24] <NCommander> ah
[11:24] <asac> mozilla isnt good at that. but i have yet to see that webkit is better ;)
[11:24] <asac> s/good/perfect/ ;)
[11:25] <NCommander> I just perfer gecko over webkit at the moment on desktops
[11:25] <NCommander> although if Presto was opened up ....
[11:25] <asac> for the log: i think mozilla does a good providing security updates for their products - even if its not exactly the kind of support we would like to see for their rendering engine (e.g. long term support)
[11:26] <NCommander> Well, Dapper using FF 2 still which is now dead upstream
[11:26] <ScottK-laptop> NCommander: 1.5
[11:26] <asac> anyway. i am happy that webkit exists. that gives us competition and puts some pressure on the gecko developers
[11:27] <asac> 1.5 branch is maintained by distros yes.
[11:28] <NCommander> ScottK-laptop: OW
[11:28] <NCommander> MY BRAIN
[11:28] <NCommander> OW
[11:28] <asac> but we do quite a good job imo. we didnt see any considerable regressions and are still on top of all security patches - which is a lot.
[11:29]  * ScottK-laptop has a Dapper desktop upon which he uses FF every day and agrees.
[11:31] <NCommander> ScottK-laptop: please backport FF3 to dapper
[11:32] <NCommander> ScottK-laptop: BTW, what were you asking me on SUbversion? (I posted a fix to correct the SSL certificate issues)
[11:34] <ScottK> NCommander: Thanks.  I'll have a look.
[12:05] <Company> mpt: ping
[12:38] <asac> ScottK: whats the status of bug 231743 ?
[12:39] <asac> ScottK: for me that patch makes sense, but i am not really up-to-date with REVU and such
[12:47] <persia> asac: The fix was applied in dput 0.9.2.33 in Debian.  Do we not want any of the other adjustments as well?
[12:48] <persia> asac: There's what looks like a useful adjustment to support PPAs as well
[12:49] <asac> persia: not sure. if we want that change i would like to upload in any case. so the contributor gets his credits. we can then still sync
[12:49] <persia> I guess.  That's the right address for REVU at least.
[12:50] <persia> asac: Note that the other address also works, so it's really a no-op.
[12:50] <persia> Personally, I think it's pointless to do this as an Ubuntu change.
[12:51] <persia> (and it's the sort of thing I count negatively when considering package uploads when doing a technical review)
[12:51] <asac> persia: right but he submitted that patch quite some time ago. and its a problem in the sponsoring process that this is now more or less outdated.
[12:52] <asac> persia: but ok.
[12:52] <persia> asac: So?  Does it make more sense to apply now?  Considering that it has *zero* impact on operations?  You can upload it if you like.
[12:52] <cjwatson> we should thank him but then do whatever the right thing is in terms of getting the actual change into Ubuntu
[12:52] <asac> ok
[12:52] <cjwatson> sometimes multiple people submit a patch and we don't necessarily always credit them all
[12:52] <persia> Also a useful question: did it make sense to apply previously?  The last two sponsors didn't think so.
[12:53] <persia> cjwatson: True.  I try to go for earliest submission for credit when sponsoring, but that's not always simple (although it simplifies reviews)
[12:53] <asac> cjwatson: right. but in this case we should have pointed him to debian in _may_ :)
[12:53] <cjwatson> oh, I agree
[12:54] <cjwatson> I just mean that we shouldn't do the wrong thing in terms of upload purely in order to be able to give credit
[12:54] <cjwatson> there are other ways to thank contributors
[12:55] <asac> i fully agree. its just that it wont show up in +packages ... which appears to be a prominent place to review contributions for MOTU membership and such :)
[12:56] <persia> asac: Well, no.  +packages is completely broken (and has been for over six months).
[12:56] <persia> Generally, one needs to review the -changes archives.
[12:56] <asac> imo that way of looking at MOTU contributions should be fixed
[12:56] <asac> but until that is happened it pops up as "i dont see many contributions of you".
[12:56] <asac> persia: ack+
[12:57] <asac> ok i will review the debian package and ask for sync if the changes are ok
[12:57] <asac> but now lunch
[12:57] <cjwatson> +packages doesn't work if changes from multiple contributors are included in one upload either. It's a completely unreliable method of gauging contributions whatever way you slice it.
[12:57] <persia> asac: I'll keep saying "I don't see many contributions" if I don't, but it's not always the tools to blame.  More generally, we've outgrown the ogre model, and no longer effectively have two teams.
[12:57] <cjwatson> and the contributor can and should defend themselves by pointing out their contributions
[12:58] <cjwatson> I don't have much sympathy for people who don't reply to that sort of thing :)
[12:58]  * ogra wonders why his SD card is sometimes mounted and sometimes ignored by gvfs
[12:58] <persia> Absolutely.  The instructions for application are to list some people who can give strong endorsements, and provide examples of work of which they are particularly proud.  This is rarely actually provided.
[12:59] <ogra> its the same card ... but seems it depends on the phase of the moon if it gets mounted or not ... dmesg shows the proper events
[13:01] <persia> ogra: I had those symptoms with an SD card under linux 2.4.  It turned out to be related to how it was unmounted last.  Could be a different problem for you, of course.
[13:01] <ogra> well, gnome-volume-manager did magae it right in former releases
[13:01] <ogra> gvfs doesnt it seems
[13:02] <ScottK> asac: Coming back in a little late, but I think sending the change to Debian and leaving it at that is the right answer.
[13:03] <ScottK> OTOH, if someone feels it's worth uploading, I wouldn't object or anything.
[13:04]  * ogra starts to suspect a polkit issue ... 
[13:05] <ogra> the "places" link is shown fine but gnome-mount doesnt do anything with it
[13:06] <seb128_> gnome-mount -b -v -d device
[13:57] <jwendell> seb128, hi! login sound is not working on intrepid. Is that a know issue?
[13:57] <seb128> hey jwendell
[13:58] <seb128> jwendell: yes, libcanberra requires the freedesktop sound theme which is not available yet
[13:58] <jwendell> ah ok
[13:58] <jwendell> seb128, will it be available in time to intrepid?
[13:58] <jwendell> do you know?
[13:59] <seb128> jwendell: no I don't know
[13:59] <seb128> debian has it packaged but it didn't get accepted there for licensing reasons
[13:59] <jwendell> hmm
[13:59] <seb128> and I've been too busy to push it to ubuntu, need to look into it again
[14:00] <jwendell> need some help?
[14:01] <seb128> jwendell: you are welcome to try to get it uploaded to ubuntu ;-)
[14:01] <jwendell> ok, will try
[14:02] <seb128> jwendell: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=486559
[14:03] <seb128> jwendell: I think one issue is that the upstream tarball has no license texts
[14:03] <psyke83> asac: hi, I got your message on the forums. Most of the activity on the forums is using guinea pigs to test the PulseAudio "PerfectSetup" configuration (bug #192888, bug #198453) and test the new RC's of Flash. Luke's doing most of the heavy lifting now that he's maintaining preview packages of PulseAudio
[14:04] <seb128> jwendell: sjoerd_ might know what were the issue for debian and if somebody is working on fixing those
[14:04] <asac> psyke83: ok. so you are already working with Luke?
[14:04] <asac> psyke83: that makes me feel better.
[14:05] <ogra> asac, i still dont get why we cant just SRU flash 10 for the libflashsupport issue
[14:05] <psyke83> asac: I haven't been active on IRC for a while (I'm part of the ubuntustudio dev team, so I have contact with Luke there), but I've been reporting bugs on Launchpad against his builds
[14:05] <asac> psyke83: whats the idea? did you find a config for the pulseaudio alsa plugin that auto fallbacks to default behaviour for kde?
[14:06] <asac> psyke83: ok. thanks for clarifying. my main concern is flash  + pulse in intrepid. so if you need backup or help there let me know - though i think that Luke is probably on top.
[14:06] <psyke83> asac: I believe Daniel had some patches pending to set this behaviour, see: https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2008-August/005301.html
[14:06] <asac> ogra: for the same reason that we still have no perfect solution in intrepid
[14:07] <asac> ogra: and also from what i understood crimsun the pulse plugin for alsa in hardy wouldnt be good enough anyway and would need fixes
[14:07] <ogra> asac, i dont get it ... according to lennart flash10 works flawless without libflashsupport, intrepid has flash 10
[14:07] <psyke83> asac: however, we can handle the situation for non-PA configs in another way, without the need for patches. We can create a wrapper script so that the PulseAudio ALSA plugins are enabled (asoundconf set-pulseaudio) only on systems where PulseAudio is actually installed
[14:08] <asac> ogra: upstream people sometimes are over optimistic and dont consider real life. we have to setup the pulse audio plugin for alsa by default to make it work perfectly
[14:08] <ogra> psyke83, have a look at ltsp, it does exactly that with an Xsession.d script
[14:08] <asac> ogra: but if we do that we break kde.
[14:08] <asac> ogra: thats why we are looking for better configs.
[14:08] <norsetto> seb128: hi seb, can I query you?
[14:08] <asac> ogra: and on top the pulse plugin in hardy appears to be not good enough
[14:09] <asac> (for the hardy backport)
[14:09] <seb128> hey norsetto, yes
[14:09] <ogra> weird that we never had a prob with pulse in ltsp in gutsy :)
[14:09] <ogra> (*with* libflashsupport)
[14:09] <psyke83> asac: have you seen the packages in my PPA? I backported the necessary parts to get PulseAudio working acceptably in Hardy (not the newest PulseAudio, but 0.9.10)
[14:10] <cjwatson> seb128 (and maybe others): does bug 202038 make sense to you?
[14:10] <asac> psyke83: lets first get all the required bits into intrepid
[14:11] <asac> afaik, this still hasnt happened (or am i wrong)?
[14:12] <asac> ogra: my story about this is that i that recent flash 9 updates probably changed the underlying infrastructure, which then made the unmaintained libflashsupport cause deadlocks/break
[14:12] <seb128> cjwatson: I guess he would like an empty gnome-keyring password by default on the livecd so users don't get prompted for gnome-keyring password the first time it's used
[14:12] <asac> and nobody can figure out how to update libflashsupport (because there is no info from adobe obviously)
[14:13] <ogra> asac, yeah, i remember now ...
[14:13] <ogra> gutsy was flash7
[14:13] <leleobhz> compiling pymedia:
[14:13] <leleobhz> In file included from audio/acodec/acodec.c:31:
[14:13] <leleobhz> audio/libavcodec/dsputil.h:485: error: static declaration of ‘lrintf’ follows non-static declaration
[14:13] <leleobhz> how can i solve this?
[14:13] <cjwatson> seb128: or perhaps memory reduction
[14:13] <asac> ogra: yeah ... flash 9 was a huge updated. from what i can tell they shipped flash 7 engine + a completely new engine (for flash 9 aka action script 3) in the same player
[14:13] <psyke83> asac: I've been investigating libflashsupport, and it cannot be fixed. If you test Adobe's original implementation of libflashsupport (OSS/ESD support), it causes crashes as well. The API seems to be buggy
[14:14] <asac> psyke83: the api isnt buggy. it just changed
[14:14] <asac> and nobody knows how it was changed
[14:14] <ogra> psyke83, well, the flash api changed
[14:14] <asac> and if its actually really ment to work
[14:14] <ogra> psyke83, libflashsupport was developed for flash 7
[14:14] <ogra> whith which is worked flawless
[14:15] <psyke83> ogra: http://labs.adobe.com/wiki/index.php/Flash_Player:Additional_Interface_Support_for_Linux
[14:15] <seb128> cjwatson: right, that too
[14:15] <psyke83> you're referring to Adobe's original implementation being flawless, right?
[14:15] <ogra> psyke83, right and the one lennart pulled into the pulse source tree
[14:16] <ogra> we used it in gutsy in ltsp when we started to use pulse on the clients and a virutal alsa device on the server side without any probs
[14:16] <ogra> but that was flash7
[14:16] <psyke83> I compiled Adobe's version of libflashsupport, and tested it *without* PulseAudio. I was still experiencing the segfaults as per bug #192888. My conclusion is that the API is buggy, and the problem isn't on the PulseAudio side
[14:16] <ogra> right
[14:16] <ogra> between flash 7 and 9 something changed
[14:17] <ogra> on the flash sie
[14:17] <psyke83> I can't speak for flash 7, but I tested the "original" library on 9 and 10 and both crash
[14:17] <ogra> libflashsupport was never changed
[14:17] <ogra> the code is identical to what they released with flash 7
[14:18] <NCommander> seb128: morning
[14:18] <ogra> lennart made some minor changes during the flash 8 times
[14:18] <psyke83> ok, when I said the API was buggy, I wasn't too clear, sorry. I meant to say that the API is exposing a bug within the Flash binary (i.e. the part we can't see)
[14:18] <seb128> hello NCommander
[14:18] <ogra> right
[14:18] <asac> psyke83: ok. lets not look back, but to the future. when do you think could we have this startup script magic for intrepid?
[14:19] <NCommander> seb128: mind retrying the gtkmm builds?
[14:19] <NCommander> (buildd retry gtkmm should do the trick)
[14:19] <asac> psyke83: saying that libflashsupport exposed a bug in flash is like saying that a binary build against gtk 1 exposes a bug in gtk 2 ;)
[14:19] <ogra> asac, having a Xsession.d script is totally trivial http://paste.ubuntu.com/47442/
[14:19] <seb128> NCommander: I did half an hour ago?
[14:19] <psyke83> asac: I'll try to get in touch with Luke and discuss this, and see if PA 0.9.12 deserves a FFe. No matter, I'll work on this script and let you know when I have news
[14:19] <NCommander> Oh
[14:19] <seb128> NCommander: and I promoted pangomm to main too, will be effective after the next publisher run and unblock the depwait
[14:20] <NCommander> GUess Launchpad being screwy, it still says "gtkmm No builds attempted"
[14:20] <asac> psyke83: not exactly the same, but you get the point. maybe its a bug in the flash player. but it could as well just that libflashsupport wasnt adapted to the new flash contract ;)
[14:20] <asac> psyke83: why do we need a new upstream release for that?
[14:20] <psyke83> asac: IMHO, we need to completely eradicate this vermin that is libflashsupport ;). That also means removing it from ia32-libs for 64bit users
[14:20] <NCommander> why is libflashsupport still needed?
[14:20] <psyke83> asac: we don't need a new upstream release, it's just that Luke may know more details than I, perhaps the new release fixes other issues
[14:20] <NCommander> I haven't needed it since upgrading to intrepid
[14:20] <cjwatson> seb128: put another way, is there likely to be any actual downside from disabling gnome-keyring on the live CD?
[14:21] <ogra> the prob is that lennart will refuse to update libflashsuport even if adobe would provide fixes since he thinks flash10 works fine out of the box
[14:21] <asac> psyke83: thats understood. but want to see it fixed for real ;)
[14:21] <cjwatson> seb128: I suppose it would be inconvenient if you were using the live CD as a glorified ssh client in a cybercafe or something
[14:21] <psyke83> ogra: flash 10 does work fine, on a properly configured PulseAudio system
[14:21] <asac> psyke83: i am sure that PA might fix other issues. but i am most concerned about flash still not having proper sound everywhere ;)
[14:21] <norsetto> NCommander: did you see bug 270698 ?
[14:21] <seb128> cjwatson: how would you disable it? I don't think it allows that at the moment
[14:21] <ogra> psyke83, right, but hardy has 9 :)
[14:22] <cjwatson> remove the dbus service? :-)
[14:22] <seb128> NCommander: no build attempted, it depwait on pangomm
[14:22] <NCommander> norsetto: no, I didn't
[14:22] <psyke83> asac: sure, I understand. We don't need to update PulseAudio to fix Flash. All that's needed is a) to remove all traces of libflashsupport, b) update ia32-libs, and c) work on a script to set the PA ALSA plugins only when it's appropriate
[14:22] <cjwatson> disable_kwallet works in much the same kind of way - it removes a file in /usr/share/services/kded/
[14:23] <NCommander> norsetto: I cleared the FTBFS, made sure it installed, but beyond making sure it installed and the binaries "worked", I didn't actually try setting up a test server
[14:23] <norsetto> NCommander: sure but pls. check it out, we might need to patch it
[14:23] <asac> psyke83: right. my only point is that the flash fix has been dragged for much too long. so whatever is required to fix that it should be treated asap. if that happens with a new upstream version or in a distinct fashion - i dont care ;)
[14:23] <seb128> cjwatson: well, I'm not sure how network-manager will react when trying to store the passphrase in gnome-keyring if gnome-keyring is not running, I didn't try that
[14:23] <NCommander> norsetto: you lost me
[14:23] <psyke83> ogra: realistically speaking, we're probably going to be forced to flash 10 on hardy. The latest RC is versioned as 10.0.12.10 which indicates the final release will be soon. As soon as there's a security advisory, Flash 10 will be a mandatory upgrade
[14:23] <seb128> cjwatson: same for evolution (or other applications using gnome-keyring)
[14:23] <asac> psyke83: can you include my nick when you approach luke on this? when i am online iwould surely like to contribute to that discussion ;)
[14:24] <norsetto> NCommander: why, busy saving the lehman guys from suicide?
[14:24] <psyke83> asac: sure
[14:24] <seb128> cjwatson: you can try, but I would not be surprised if those display errors if gnome-keyring is not working
[14:24] <NCommander> norsetto: now you REALLY lost me
[14:24] <ogra> psyke83, right, thats been my intro question :) "<ogra> asac, i still dont get why we cant just SRU flash 10 for the libflashsupport issue "
[14:24] <cjwatson> seb128: hmm, ok, fair enough
[14:24] <cjwatson> thanks for the feedback
[14:24]  * norsetto looks around for NCommander
[14:24] <NCommander> ogra: when we tried that via backports, it was bad, really really bad
[14:25] <NCommander> norsetto: huh?
[14:25] <ogra> NCommander, because you made ral bad mistakes
[14:25] <norsetto> NCommander: ah ok, I found you. What is it that you don't get?
[14:25] <NCommander> ogra: I didn't do that backport, I just know that it ended badly
[14:25] <NCommander> norsetto: I don't get by patching, if all you need to do is add a configure flag
[14:25] <ogra> NCommander, libflashsupport *doesnt* work with libflashsupport but whoever backported it left the dependency in
[14:26] <ogra> err
[14:26] <psyke83> ogra: it's not so simple. If you remove libflashsupport in Hardy, then it will rely on the PA ALSA plugins, and alsa-lib/alsa-plugin would need an update due to bugs in Hardy's versions
[14:26]  * NCommander blinks
[14:26] <ogra> *doesnt* work with flash10
[14:26]  * NCommander decides /dev/coffee is needed and lights the stove
[14:26] <ogra> psyke83, right
[14:26] <psyke83> ogra, in my PPA I backported alsa-lib and alsa-plugins, and it all works fine for Hardy users
[14:27] <ogra> psyke83, but leaving the dep to libflashsupport in the flash10 backport was a mistake nontheless
[14:27] <psyke83> in fact they were 1.0.16, not the latest version in Intrepid
[14:27] <ogra> and the base of all subsequent evil :)
[14:27] <NCommander> norsetto: what do you need for me?
[14:27] <asac> ogra: right. but that backport was removed?
[14:27] <norsetto> NCommander: oh dear ... patching like "adding the configure flags" not as "add your own patch system and add a patch in debian/patches"
[14:27] <ogra> asac, yep
[14:27] <psyke83> ogra, I believe it's not a dep anymore, it's libflashsupport | libasound2-plugins
[14:27] <asac> afair it was an accident that that was ever pushed out
[14:27] <NCommander> norsetto: yeah, /dev/coffee is needed
[14:27] <ogra> asac, right
[14:27] <NCommander> norsetto: As I said before, I simply cleared the FTBFS, I didn't make sure the server worked beyond making sure it installed
[14:28] <ogra> asac, though i could have told the backporter about it ... my name sits as packagedr in the libflashsupport package and it would have been one mail to ask :)
[14:28] <norsetto> NCommander: I understand that, but you surely are in a better position to follow that bug report up, or not?
[14:29] <NCommander> norsetto: OH!
[14:29] <NCommander> norsetto: now I get it
[14:29]  * NCommander is right now having a ENOCAFFIENE state of mind
[14:30] <asac> ogra: i could have told him too
[14:30] <ogra> yeah
[14:30] <ScottK> asac: Accident is probably too nice a way to put it.  It was certainly not the best considered backports decision I've ever made.
[14:30] <asac> ogra: it just happens that it was approved by accident
[14:30] <ogra> yeah
[14:30] <asac> ScottK: well. shit happens ... i dont think its a big problem ;)
[14:31] <ScottK> Yeah.
[14:31] <asac> but it might indicate that there is room for improvement on how we ack backports
[14:31] <ogra> nah
[14:31] <ogra> and its over anyway :)
[14:31] <NCommander> norsetto: http://qdb.us/224378 - I'm reminded of this for some reason
[14:33] <NCommander> what the heck is with Freenode this week, its been netsplitting like nuts it seems.
[14:34] <norsetto> NCommander: hehe
[14:34] <NCommander> I've done that actually
[14:34] <NCommander> Not my smartest moment ever
[14:38] <NCommander> Oh man
[14:39] <NCommander> Instead of coffee I got sludge ;.;
[14:39]  * Hobbsee covers NCommander in green sludge.
[14:40]  * NCommander pours his coffee on Hobbsee and watches her desolve
[14:40] <Hobbsee> ewww
[14:40] <psyke83> asac: I mentioned that crimsun had a patch for the alsa-plugins to fallback to ALSA, but I see in the bazaar revisions that Luke is working on a better approach. I'll have a chat with him and see what the progress is
[14:41] <radix> kirkland: would you have a moment to discuss bug #270131?
[14:42] <radix> I'm not sure about the hanging part, but I'm trying to debug the bit about 'ln' failing
[14:42] <asac> psyke83: good. please summon me in that discussion (by naming my nick) so i lurk or read the backlog
[14:44] <psyke83> TheMuso: asac wants to know how you're going to deal with the ALSA fallback for the PA ALSA plugins ;)
[14:44] <NCommander> Hobbsee: coffee is never ewww. Its the life blood of millions
[14:44] <psyke83> asac: I think Luke's pretty busy, so I'll fire off an e-mail and forward anything I get to you
[14:44] <asac> psyke83: well. that doesnt help me if the discussion doesnt just happen now ;)
[14:44] <psyke83> hehe
[14:46] <asac> psyke83: feel free to CC me in that email: asac@ubuntu.com so you dont need to forward. thanks!
[14:46] <psyke83> sure thing
[14:46] <NCommander> ScottK: got a chance to look at the SVN backport patch?
[14:47] <ScottK> NCommander: Not yet.  $ELDEST_CHILD missed the bus today so I had to drive her.  I'm kind of behind.
[14:47] <NCommander> ScottK: sounds like fun. At least you didn't get sludge for coffee
[14:48] <ScottK> NCommander: I was in the Navy and drank "Still here from yesterday, but it's the midwatch and I'm REALLY tired" coffee for years.  All the coffed I have now is good.
[14:49] <NCommander> ScottK: rofl. That's not too bad
[14:49] <NCommander> ScottK: I believe my coffee would count as a chemical weapon in some countries
[14:50] <ScottK-laptop> Unless you've experienced it, you don't know.
[14:50]  * NCommander poours a cup of it on the ground and watches it eat right through the bottom of this channel
[14:50] <NCommander> ScottK-laptop: I've had the coffee machine hasn't been cleaned since the turn of the century coffee a few nights ago
[14:50] <StevenK> NCommander: Then how does it stay in the cup?
[14:50] <NCommander> StevenK: lead cup
[14:51] <NCommander> (638 was our command vechile, and its coffee maker is probably older than you are)
[14:51] <StevenK> Bit hot to hold
[14:51] <ScottK-laptop> NCommander: If you want to play that game, I have a coffee cup here that I started 'seasoning' in 1988 (I am not making this up).
[14:51] <NCommander> Crap, that coffee is older than I am
[14:52] <NCommander> ScottK-laptop: I had refridgated coffee I forgot about, and when I found it when we cleaned the fridge out when we sold the house ...
[14:52] <NCommander> Well
[14:52] <NCommander> I didn't know coffee could mutate like that
[14:52] <NCommander> The worst though is crunchy coffee
[14:54] <ogra> well, hot water can revive it
[14:54] <NCommander> ogra: it was hot, crunchy, coffee
[14:54] <ogra> oh
[14:54] <NCommander> Our MTO forgot to put a filter in I think
[14:54] <NCommander> (its the only way I can guess it could happen)
[14:54] <johanbr> At my previous place of work, people would sometimes come in on Monday morning and drink the coffee that had been left on the pot over the weekend.
[14:54] <NCommander> It was corsiacly ground, so it floated
[14:55]  * NCommander is always amazed on how bad some coffee machines get
[14:55] <NCommander> When we cleaned the one out at the training center, we used vinger, and we got a black tar :-/
[14:56] <NCommander> I think we ended up using something the chemical equivelent to battery acid to clean that machine out
[14:56] <ScottK-laptop> Cleaning is always a mistake.  Then it has to be properly seasoned again and that can take ages.
[14:57] <NCommander> ScottK-laptop: when your coffee comes out as tar vs a fluid, I think its worth cleaning
[14:57] <persia> NCommander: No, that's when it's worth drinking
[14:58] <ogra> now you guys made me thursty ...
[14:58]  * ScottK-laptop high fives persia.
[14:58] <NCommander> you like tarry coffee?
[14:58]  * ogra makes some fresh coffee
[14:58]  * StevenK doesn't drink coffee
[14:58]  * NCommander used the emergency instant since the machine is broken it seems
[14:58] <ogra> StevenK, how do you manage these 18h days without coffee o_O
[14:58] <StevenK> ogra: Tea
[14:58]  * broonie always has a backup cafetiere for these purposes.
[14:58] <NCommander> "In case of no coffee, break glass"
[14:58] <ogra> ah
[14:59] <StevenK> ogra: And powdered tooth enenemal
[14:59] <ogra> french press FTW
[14:59] <persia> NCommander: I no longer drink coffee.  When I was a coffee drinker, my favorite preparation was to make 4 cups of espresso, boil for about 10 minutes to get rid of the excess water, and mix 2 parts with one of condensed milk.
[14:59] <NCommander> persia: o_o;. You have no taste
[14:59] <NCommander> ANd I mean that as taste buds
[14:59] <ogra> french press with freshly grinded espresso though :)
[15:00] <StevenK> persia: And how did you get a stronger hit? With a needle?
[15:00] <persia> NCommander: coffee doesn't affect the taste buds: it's a smell thing.
[15:00] <ogra> so it still stays liquid
[15:00] <NCommander> persia: after drinking that, I would be suprised if you have any taste buds left
[15:00] <ogra> having to use a spoon and to chew coffee simply isnt the right thing
[15:02] <jdstrand> mvo: hi!
[15:02] <NCommander> persia: maybe we should snort coffee vs drink it ...
[15:02] <persia> StevenK: boil strong tea.  Mix with NaHCO3.  bake.  Boild again.  Cool.  Add dichroromethane.  mix.  Dry thouroughly.  Filter with Na2SO4.  aspirate.  finely powder.  smell gingerly.
[15:03] <persia> ogra: It's very popular that way in some parts of the world.
[15:03]  * _MMA_ laughs as #ubuntu-devel descends into coffee madness. :)
[15:03] <NCommander> It's on topic. ubuntu development is driven by coffee
[15:03] <persia> No.  It's not really ontopic :p
[15:03] <jdstrand> mvo: so I have a situation where I would like to do basic authentication with https via apt. I installed apt-transport-https, but I get a 401. I check through the docs, but don't see how to set the username/password in apt.conf (this isn't a proxy). is this possible?
[15:04] <StevenK> MOTUs and core-devs turn coffee/tea into packages and bug fixes.
[15:05] <NCommander> StevenK: now if we could only some how turn packages and bugfixes into coffee, we'd have a self-sustaining ecosystem of coffee
[15:05] <StevenK> Haha
[15:06] <NCommander> I figure if we could make Ubuntu coffee, we could simply close bug #1 by converting the masses to our uber and free coffee
[15:10] <mvo> jdstrand: yes, but you need to add the password into the sources.list (like deb http://user:pass@foo.com/)
[15:10]  * jdstrand slaps head
[15:10] <mvo> jdstrand: it may also complain (in intrepid) about the certificatio (just like firefox) if you only have a self signed one
[15:10] <jdstrand> mvo: cool-- thanks!
[15:10]  * jdstrand nods
[15:11] <mvo> jdstrand: I would like to overcome this limitation (the password in sources.list one) but hadn't had time yet
[15:11] <jdstrand> mvo: but it seems there are various knobs for that
[15:11] <mvo> indeed
[15:11] <jdstrand> mvo: I can chmod 600 it for now
[15:11]  * mvo nods
[15:12] <ion_> summon Keybuk
[15:13] <ion_> Not seb128, i said Keybuk.
[15:13] <seb128> ion_: ?
[15:13] <ogra> ion_, focus better then
[15:13] <ion_> < ion_> summon Keybuk  -!- seb128 [n=seb128@ubuntu/member/seb128] has joined #ubuntu-devel ;-)
[15:13] <seb128> ah
[15:14] <seb128> yeah, try again
[15:14] <NCommander> Nope
[15:14] <NCommander> THat just made people leave
[15:14] <NCommander> Try again
[15:15] <persia> Umm.  Shouldn't that mention the soft-freeze for Alpha-6?
[15:15] <NCommander> hrm
[15:16] <persia> Maybe s/alpha-5 released/Soft freeze for Alpha-6/ ?
[15:16] <persia> THat works.
[15:16] <NCommander> I just wanted to remove the buildds were down message
[15:16] <NCommander> since it was old
[15:17] <ogra> is the freeze in effec already ?
[15:17] <jdstrand> mvo: is the plan something along the lines of Acquire::http::<server> "https://<user>:<pass>@<server>"; or some such?
[15:17] <jdstrand> (in apt.conf or apt.conf.d)
[15:17] <persia> NCommander: https://launchpad.net/+builds shows several buildds down
[15:18] <persia> (well, at least two)
[15:18] <ScottK-laptop> persia: Bet the message was from when they were all dead due to bad upload.
[15:18] <ScottK-laptop> Bet/But
[15:18] <jdstrand> mvo: btw, user@pass in sources.list worked just great
[15:18] <persia> Ah.  Yeah.  That was sorted.  There seems to have been a marked reduction in PPA buildds as well, but packages are being built.
[15:22] <cjwatson> ScottK-laptop: no, lots of buildds were taken down in order to move them between datacentres
[15:22] <mvo> jdstrand: I was thinking of something like /etc/apt/netrc - however I'm not sure if its flexible enough
[15:22] <cjwatson> that was yesterday morning though, so yes it's fine to remove that message
[15:22] <ScottK-laptop> Ah.  OK.  I thought it was from the gcc 4.3 problem last weekend.
[15:39] <mathiaz> slangasek: have you started to the Alpha6 Release notes wiki page ?
[15:39] <mathiaz> kirkland: ^^
[15:39] <kirkland> mathiaz: thx ;-)
[15:44] <mathiaz> kirkland: has update-motd been accepted into main ?
[15:44] <kirkland> mathiaz: waiting on response from pitti
[15:44] <kirkland> mathiaz: i believe that i have addressed his concerns in a sequence of uploads yesterday
[15:45] <kirkland> mathiaz: i vetted my changes by kees, and he seemed to it looked okay
[15:45] <kirkland> mathiaz: you're welcome to review them too
[15:45] <kirkland> mathiaz: see the MIR bug for update-motd for detailed explanation of changes
[15:45] <mathiaz> kirkland: ok - landscape-client doesn't currently installs from the -server iso because update-motd is missing.
[15:46] <kirkland> mathiaz: i understand this....  and i'd push update-motd if i were an archive admin
[15:46] <kirkland> mathiaz: :-)  but que sera sera
[15:46] <kirkland> mathiaz: i don't see pitti or doko online at the moment
[15:48] <persia> kirkland: mathiaz: https://wiki.ubuntu.com/ArchiveAdministration lists the duty schedules for the archive admins.  Today is Tuesday.
[15:48] <kirkland> persia: thanks...
[15:48]  * kirkland wonders if Riddell is around :-)
[15:49] <persia> Now you're looking for the right person :)
[15:49] <mathiaz> persia: The MIR hasn't been accepted yet.
[15:50] <mathiaz> persia: IIUC we first need to get that step done and then we can bring in the archive admin team :D
[15:50] <persia> mathiaz: Oh, so it's not NEW, but promotion?  Yeah, that's a smaller set of people.
[15:52] <mathiaz> persia: yes - a set composed of 2 people that are both not around today.
[15:52] <kirkland> persia: right, pitti raised some concerns, i think i've addressed them, would like him to review my changes and perhaps approve the MIR this time
[15:53] <persia> Sorry.  I misunderstood.  I hope the link will be useful later, although pitti usually does the promotion when approving the MIR.
[15:55] <cjwatson> NCommander: I can't reproduce gdm prompting at login on Xubuntu CDs. which boot option are you using?
[15:55] <cjwatson> NCommander: and is this reproducible with current dailies?
[15:56] <NCommander> cjwatson: default options on amd64, haven't had time to check the dailies
[16:34] <tseliot> superm1: are you planning to change the names of symlinks such as /var/lib/dkms/nvidia/kernel-2.6.27-2-generic-i686 in DKMS?
[16:35] <superm1> changing the name of symlinks?
[16:35] <superm1> tseliot, do you want to elaborate?
[16:36] <tseliot> superm1: when you build and install a module for a certain kernel a symlink such as kernel-$(uname -r)-i686 is created, right?
[16:37] <tseliot> in the nvidia or fglrx directory
[16:37] <superm1> right
[16:37] <tseliot> superm1: you're not planning to change the name scheme of these symlinks, are you?
[16:38] <tseliot> just to be sure
[16:38] <superm1> not anytime soon, no
[16:38] <tseliot> ok
[16:40] <tseliot> superm1: there are still users who have several directories in /var/lib/dkms/nvidia from older drivers because of that bug (which I fixed) in the postinst
[16:40] <tseliot> and I want to make sure that such directories are removed when a new driver is installed
[16:41] <superm1> tseliot, well make sure that you check for upgrades only from broken versions when you are upgrading
[16:41] <superm1> tseliot, so that this problem will conditionally eventually go away then
[16:42] <tseliot> superm1: yes, sure but I would like to avoid deleting those symlinks. This is why I asked you
[16:42] <tseliot> about the name schemes
[16:44] <tseliot> superm1: oh, and did you send my patch to NVIDIA?
[16:45] <tseliot> (nvidia-settings)
[16:45] <superm1> tseliot, i brought it up to them, they haven't responded to me about it yet though
[16:45] <tseliot> superm1: ok, thanks
[17:12] <didrocks> ScottK: around?
[17:13] <NCommander> hey lamont, are you around?
[17:13] <NCommander> or anyone with a hppa box?
[17:13] <ScottK-laptop> didrocks: What's up?
[17:14] <didrocks> ScottK-laptop: I am trying to backport ufw to hardy (from a request on -server ML). For that, I had to backport iptables
[17:14] <didrocks> I want just to be sure I have followed the process described at https://help.ubuntu.com/community/UbuntuBackports
[17:14] <didrocks> current state is:
[17:15] <didrocks> - I had to make a trick so that iptables compiles (and run) on hardy
[17:15] <didrocks> - no change has to be done for ufw
[17:15] <didrocks> the associated bug report is at https://bugs.launchpad.net/ubuntu/+source/iptables/+bug/268931
[17:15] <NCommander> didrocks: I'm not sure iptables is elliable for backports, thats a massive change.
[17:16] <mkrufky> superm1: hello
[17:16] <didrocks> NCommander: yes, the backport was not straight, that's why I want to know if it's possible or not :)
[17:16] <ScottK-laptop> didrocks: I'm not at all comfortable with backporting iptables.  At the very least it's have to be tested with it's rdepends and the list is long.
[17:16] <superm1> hi mkrufky
[17:16] <mkrufky> superm1: bug 199398 ....  i think enough time went by now... can we upload our w_scan package?  ;-)
[17:17] <mkrufky> superm1: especially considering the recent activity there... since you already did the packaging, may as well use it somewhere...
[17:17] <didrocks> ScottK-laptop: for sure, is it possible to have some king of automation to test with them?
[17:17] <NCommander> does anyone have hppa hardware?
[17:17] <ScottK-laptop> didrocks: If you have such a thing, but AFAIK nothing currently exists.
[17:17] <mkrufky> superm1: and FYI, i uploaded the latest version (with ATSC support) to bug 258479
[17:17] <mkrufky> :-/ ... thats annoying
[17:18] <didrocks> ScottK-laptop: ok, so, atm, I can build it in my ppa and try to ask for volonteers ?
[17:19] <soren> Do we have a preferred python templating engine? I don't see any in main at this point.
[17:19] <NCommander> didrocks: we usually use the backport testing PPA
[17:19] <ScottK-laptop> didrocks: Sure, but be sure to document what's tested and how so when I ask you a lot of questions, you'll have answers.
[17:20] <ScottK-laptop> soren: I don't think we do.  In #debian-python I find a lot of pylons fans.
[17:20] <soren> Cheetah seems to be the canonical (no pun intended) python templating engine.
[17:20] <didrocks> ScottK-laptop: ok, I will take some notes
[17:20] <soren> That's more of web framework, isn't it?
[17:20] <ScottK-laptop> I guess.
[17:20]  * soren digs Django for that sort of thing.
[17:20] <ScottK-laptop> Honestly this Web 2.0 stuff hurts my brain.
[17:21]  * soren chuckles
[17:21]  * ScottK-laptop just updated his web site to actually use CSS, so is somewhat behind.
[17:21] <soren> I'm not actually using it for any web related stuff
[17:21] <ScottK-laptop> No idea then.
[17:21] <soren> Err... I mean..
[17:21] <soren>  I use Django for web related stuff.
[17:21] <soren> ...but I don't intend to use the templating engine for web stuff. It's for VMBuilder.
[17:35] <slangasek> mathiaz: it's a revolving document, https://wiki.ubuntu.com/IntrepidIbex/TechnicalOverview
[17:40] <ScottK-laptop> NCommander: On your svn backport fix, the bug says pending test results.  Do you have any test results?
[17:42] <NCommander> ScottK-laptop: hrm. I forgot I said that
[17:42] <NCommander> none of the people from the backports bug have posted anything about the tests
[17:42] <ScottK-laptop> OK.  Absent some test results I'm a bit reluctant to proceed.
[17:54] <Riddell> kirkland: hi
[17:54] <mihai81> hi, how can someone get involved in ubuntu development
[17:55] <kirkland> Riddell: hiya, i was pinging you earlier, but I think I'm still waiting on pitti to approve the MIR ;-)
[17:58] <persia> mihai81: There are lots of ways.  I think the simplest is to find a bug on LP that you can fix, fix it, and ask lots of questions on #ubuntu-motu until it gets released.  Repeat a few times, and you'll be an expert.
[17:58] <slangasek> Riddell, ScottK-laptop: are either of you looking into the kdepim/kdesdk uninstallables by chance?
[18:00] <ScottK> No, but I suppose I could look at kdepim since I TIL.  What was uninstallable?
[18:00] <slangasek> http://people.ubuntu.com/~ubuntu-archive/testing/intrepid_probs.html
[18:00]  * ScottK looks
[18:00] <slangasek> akonadi-kde
[18:01] <ScottK-laptop> OK.
[18:02] <ScottK> slangasek: I'll look at mail-spf-perl too.
[18:04] <slangasek> great, thanks
[18:11] <ScottK> slangasek: spf-tools-perl should be demoted to Universe.  It's not needed in Main.
[18:12] <Chipzz> I was just wondering
[18:13] <Chipzz> I was setting up an SSL cert
[18:13] <Chipzz> and different packages manage SSL in diferent ways
[18:13] <Chipzz> dovecot apparently autogenerates a key
[18:14] <Chipzz> other packages do different things
[18:14] <ScottK> slangasek: For libmail-spf-perl, it looks like I MIR'ed the wrong Perl module.  I wanted libnetaddr-ip-perl and I did libnet-ip-perl.  Urgh.
[18:14] <slangasek> heh
[18:14] <slangasek> ok, I'll demote spf-tools-perl back out, at least
[18:15] <Chipzz> maybe a package ca-something could be made, which has some debconf questions; ie do you want to set up a self-signed CA, or do you want to generate a CSR
[18:15] <Chipzz> which sets things up accordingly
[18:15] <Chipzz> and other packages (like dovecot) could depend on that package, to generate the needed SSL certs automatically
[18:16] <\sh> pitti: why do I always get "can't find source blabla" when doing BUILD=0 ./fetch-and-build for ia32-libs...and always on different packages
[18:16] <slangasek> Chipzz: I believe you're describing the ssl-cert package
[18:16] <Chipzz> heh
[18:16]  * Chipzz slaps self
[18:17] <ScottK> slangasek: For kdepim, akonadi-server needs to get into Main.  The source package, akonadi is already in Main.
[18:17] <slangasek> ScottK: hey, sounds like an easy fix; promoting, thanks
[18:18] <Chipzz> slangasek: anyway, there would be some bugs in those packages for not using it?
[18:18] <slangasek> Chipzz: yes
[18:20] <Chipzz> slangasek: ah yes, I see where the confusion comes from; even when it's generated through debconf, the name still is ssl-cert-snakeoil.pem
[18:21] <ScottK> For kdepim that looks like the only issue.
[18:22] <slangasek> Chipzz: well, it's still a self-signed cert with no PKI, so it is sill snake oil unless you do something more with it ;)
[18:23] <Chipzz> slangasek: ah I see what the issue is:
[18:23] <Chipzz> slangasek:
[18:23] <Chipzz> # grep make-ssl-cert /var/lib/dpkg/info/ssl-cert.postinst
[18:24] <Chipzz> make-ssl-cert generate-default-snakeoil
[18:24] <Chipzz> so installing the ssl-cert package will always generate the snakeoil cert
[18:24] <Chipzz> it never even asks if you want to supply the information yourself
[18:24] <Chipzz> which is why I didn't know about it
[18:25] <slangasek> yes, it needs to be able to work non-interactively by default
[18:26] <Chipzz> why not use a debconf question with lower priority then?
[18:27] <Chipzz> low enough so by default it will generate the snakeoil cert, but for admins who set the debconf threshold lower, they get asked if they want to generate a cert themselves?
[18:27] <Chipzz> s/threshold/priority/
[18:30] <slangasek> Chipzz: isn't that what it does?
[18:32] <\sh> pitti: forget it..*grmpf* libltdl3 is now libltdl7
[18:33] <mathiaz> slangasek: I'm going to upload the landscape-client package split.
[18:34] <slangasek> mathiaz: o
[18:34] <slangasek> k
[18:37] <Chipzz> slangasek: no, the postinst unconditionally generates the snakeoil cert
[18:38] <Chipzz> 19:23 < Chipzz> # grep make-ssl-cert /var/lib/dpkg/info/ssl-cert.postinst
[18:38] <Chipzz> 19:23 < Chipzz> make-ssl-cert generate-default-snakeoil
[18:40] <slangasek> Chipzz: if you look inside make-ssl-cert, it uses debconf to populate the contents of the certificate
[18:41] <slangasek> so I don't see that we need to support asking the user if they want to generate a cert themselves; if we can ask them what to put in the cert instead, that should be just as good
[18:42] <Chipzz> maybe print a message along the lines of: "a default certificate was generated for you; if you want to generate one yourself, run make-ssl-cert manually"?
[18:45] <persia> No.  Non-mediated output is bad, and for many frontends, the user won't even see it.
[18:45] <fbond> What are the odds we could sync python-django 1.0-1 from sid for intrepid?
[18:50] <ScottK-laptop> fbond: There's already an approved Freeze Exception, but there is an existing Ubuntu diff that needs to be merged.  Just waiting on that to get done.
[18:52] <fbond> ScottK-laptop: thanks.  Who's doing the merging?
[18:52] <ScottK-laptop> There's an open bug on it.  I don't recall, but it's in the bug.
[18:52] <fbond> Man will I be happy to have a stable Django API in the repos... :)
[18:57] <luisbg> hello all
[18:58] <luisbg> I have a problem with how packages get installed in a group
[18:58] <luisbg> I have a package that needs to be installed after an other one (some dpkg-divert involved with a file)
[18:58] <luisbg> but if both packages are installed at the same time (apt-get install A B) which happens at install time of ubuntu studio
[18:58] <luisbg> it won't work properly
[18:58] <luisbg> :(
[18:58] <luisbg> [18:56:18] <luisbg> it is about the menu, first the package gnome-menus needs to  be installed
[18:58] <luisbg> [18:56:41] <luisbg> and then the package ubuntustudio-menu that moves the original menu to .orig and replaces it with the derivative done menu
[18:58] <luisbg> if both packages are installed at the same time only the studio menu is installed (and no .orig)
[18:59] <luisbg> this means that if a user removes the studio-menu file, the system has no menu
[18:59] <luisbg> which is not good :(
[18:59] <ogra> Keybuk is alive \o/
[19:00] <azeem> luisbg: didn't you just say the same in #ubuntu-motu?
[19:01] <Chipzz> persia: OTOH, ssh shows the output of generating keys (so you get an indication a key is being generated), while arguably it is less usefull there (there won't be many cases where you will want to generate the ssh keys with different options); so maybe having showing the openssl output so you know a key is getting generated would be a compromise?
[19:01] <Chipzz> s/having //
[19:01] <luisbg> azeem: I was then recommened to ask in here
[19:01] <persia> Chipzz: Maybe.  Having it display through mediated output would be better.
[19:02] <Keybuk> ogra: at last! :)
[19:02] <Keybuk> ogra: it has *not* been a good week
[19:03] <ogra> luisbg, have a look at edubuntu-menus how you can do that without diverting
[19:03] <ogra> Keybuk, at least you are alive :)
[19:03] <Keybuk> ogra: I am alive *and* in Portland
[19:03] <ogra> oh
[19:03] <luisbg> ogra: let me check
[19:04] <Nicke>  
[19:05] <ogra> luisbg, though persia might tell you that setting XDG_CONFIG_DIRS is evil, it is surely classes better than using dpkg-divert
[19:05] <ogra> since as you see in your current setup it breaks
[19:05] <ogra> :)
[19:06] <luisbg> ogra: so you change where XDG_CONFIG_DIRS looks for the menu?
[19:06] <Chipzz> persia: let me put it this way. I have been using debian for over 10 years, and though I don't want to sound arrogant, I consider myself highly proficient in debian (I know a lot of stuff the average user doesn't know about). If even I missed it, how will other people find out?
[19:06] <persia> Yeah, well, I wasn't advocating dpkg-diverts instead of using XDG_CONFIG_DIRS, only using DefaultMergeDirs when it was possible.
[19:06] <persia> Chipzz: Right, which is why I suggest *mediated* output.  That way it actually displays in a front-end.
[19:07] <persia> (or rather, can be displayed in a front-end).
[19:07] <persia> If you just echo something, almost nobody will see it.
[19:07] <Chipzz> slangasek just said he didn't want that
[19:08] <ogra> luisbg, yeah, install a file in /etc/X11/Xsession.d that sets XDG_CONFIG_DIRS="/usr/share/ubuntustudio" and put your .menu file in tht dir
[19:08] <luisbg> ogra: we use to do that and we switched for dpkg-diverts
[19:08] <luisbg> but now it isnt working
[19:08] <luisbg> so I do know that is a possible solution
[19:08] <luisbg> ogra: thanks :)
[19:08] <Chipzz> 19:40 < slangasek> Chipzz: if you look inside make-ssl-cert, it uses debconf to populate the contents of the certificate
[19:08] <Chipzz> 19:41 < slangasek> so I don't see that we need to support asking the user if they want to generate a cert themselves; if we can ask them what to put in the cert instead,  that should be just as good
[19:09] <slangasek> persia: you don't want debconf for this because it shouldn't be interactive; the frontends should already capture the stdout/stderr output from package installation, then it's up to the frontend to decide what to display by default...
[19:09] <ogra> luisbg, well, if that doesnt work xgd is broken (or your .menu file is incorrect) :)
[19:09] <ogra> *xdg
[19:09] <persia> Chipzz: I read that as not wanting a question to be asked, rather than not wanting information to be available.
[19:09] <luisbg> ogra: :)
[19:09] <persia> slangasek: Well, OK.  Most of the front-ends don't display anything by default :)
[19:10] <Chipzz> anyay, I don't want to make a big fuss about it; I know about it now; if you guys decide it should stay the way it is now, that's fine with me too. I just think it should be more discoverable.
[19:10] <slangasek> even a debconf 'error' would interrupt the install process
[19:11] <ogra> a note wouldnt ...
[19:11] <persia> But isn't there a NEWS mechanism that lets one inform the user of stuff, which could be used to report that a certificate was generated?
[19:11] <persia> Or a note?
[19:12] <ogra> and you ususally have a frontend ... only if you force passthrough or noninteractive you dont
[19:12] <Chipzz> (in my case it wouldn't have helped me much anyway, since I needed a CSR, and the debconf way doesn't generate one, it just generates a cert)
[19:13] <slangasek> ogra: you mean the deprecated debconf note, that is approximately equivalent to an error except sometimes it will send you email instead?
[19:14] <slangasek> which the user then won't receive...
[19:14] <ogra> slangasek, bah another of the policy police ... :)
[19:14] <Keybuk> slangasek: don't suppose you know how to change the timezone? :p
[19:14] <Chipzz> but this actually brings me to another question: how do the generated certs get used anyway?
[19:14] <slangasek> Keybuk: petition Congress?
[19:14] <Keybuk> slangasek: just in Ubuntu ;)
[19:14] <slangasek> Keybuk: dpkg-reconfigure tzdata, isn't it?
[19:14] <Keybuk> let me try that
[19:14] <Keybuk> ah, yes, that works
[19:14] <Keybuk> thanks
[19:15] <slangasek> sure
[19:15]  * Keybuk remembered there being a tzconfig or something, but that seems to have gone
[19:15] <slangasek> yep
[19:15] <Keybuk> why did that go?
[19:15] <Chipzz> since if I manually generate one, it can end up anywhere, for example outside of /etc/ssl, and I don't see how packages using ssl-cert to generate certs would find the arbitrary name I give it anyway?
[19:15] <ogra> slangasek, i havent understood why 'note' was deprecated yet ... if it was buggy sending mails that should have been fixed ... but note is a valid thing imho
[19:16] <slangasek> Keybuk: because tzconfig wasn't adequately debconf-y; now we have something that's too debconf-y instead
[19:16] <slangasek> ogra: note was deprecated because package maintainers were abusing it to chat with users
[19:17] <ogra> teach them :)
[19:17] <ogra> dont remove the tool
[19:17] <Keybuk> slangasek: I thought debconf *wasn't* a configuration database?
[19:17] <slangasek> there is no legitimate use of the tool that isn't handled by errors
[19:17] <Keybuk> ie. it wasn't a tool for modifying /etc :p
[19:17] <Keybuk> is debconf now becoming a yast-a-like?
[19:17] <ogra> slangasek, right, but an error will stop my install process
[19:17] <slangasek> Keybuk: no one ever said it's not a tool for modifying /etc
[19:18]  * Keybuk is sure joey once did, very loudly
[19:18] <slangasek> Keybuk: debconf is not a *registry*, i.e., you're not allowed to treat the contents of /var/cache/debconf as authoritative
[19:18] <slangasek> was that in between Manoj's exclamations that policy is not a stick to beat people with? :)
[19:19] <Keybuk> I've frequently wanted a stick to beat Manoj with
[19:19] <persia> policy :)
[19:19] <slangasek> tsk, elder abuse
[19:20] <ogra> persia, yeah, so policy makes ipossible to do what you wanted with the cert
[19:20] <ogra> *impossible
[19:20] <Keybuk> slangasek: it doesn't have to be elder, any kind of wood would work
[19:20] <persia> ogra: I know, which makes translations hard, which makes me want to deny Chipzz's valid desire.
[19:21] <ogra> because policy writers want to parent devs
[19:21] <persia> ogra: no, that's not why.
[19:21] <ogra> sure
 ogra: note was deprecated because package maintainers were abusing it to chat with users
[19:21] <ogra> thast totally not a valid reason imho
[19:22] <azeem> "Hi User, if you meet ogra, please tell him he still owes 5 bucks, thx"
[19:22] <slangasek> if "what you wanted" was "interrupt the install and leave the user with a message sitting on their desktop asking them to click through acknowledging that an ssl cert was created", then yes, policy doesn't allow that
[19:22] <slangasek> ogra: you seem to be missing the point that 'error' replaces 'note' for all the legitimate use cases :P
[19:22] <ogra> azeem, yeah, good idea ... we should add that as default to cdbs :)
[19:22] <persia> slangasek: No.  I just wanted a mediated exchange so it wasn't restricted to English and wasn't lost in stdout.
[19:22] <ogra> slangasek, error stops my install process
[19:22] <slangasek> ogra: so did a note!
[19:23] <ogra> it shouldnt
[19:23] <persia> I specifically didn't want to bother the user unless they were interested.
[19:23] <slangasek> except in cases where it was emailed to you instead, which is equivalent to /dev/null for default Ubuntu
[19:23] <ogra> it is a note about something the user should actually be made aware of
[19:23] <ogra> and it should work like that
[19:23] <ogra> error DTRT for an error
[19:23] <jcristau> get debian/NEWS translated
[19:24] <slangasek> then what you're asking for is not what the debconf note interface /ever/ did, so I don't see why you care that it was deprecated
[19:24] <ogra> well, it should be fixed instead of being dropped :)
[19:24] <Chipzz> couldn't notes just get queue'd, and shown in one batch at the end of the installation?
[19:25] <Chipzz> (I know technically that would be hard I suppose, but just as an idea)
[19:26] <ogra> well, ideally you shouldnt have notes ...
[19:26] <ogra> but there are cases where they make sense
[19:26] <Chipzz> lets say you install dovecot which requires ssl-cert
[19:26] <Chipzz> all installation would just go ahead
[19:27] <slangasek> NEWS.Debian, + apt-listchanges --which=news, + the i18n interface for NEWS.Debian that I never got around to writing would be a better solution for displaying messages unconditionally to the user
[19:27] <Chipzz> and you would end up with a screen which says: "Hi, I installed dovecot successfully; ALSO, this is a note from ssl-cert..."
[19:27] <Chipzz> slangasek: NEWS.Debian can't be translated though
[19:27] <slangasek> Chipzz: ... "+ the i18n interface for NEWS.Debian that I never got around to writing"
[19:28] <Chipzz> ah
[19:28] <ogra> :)
[19:28] <persia> Chipzz: It just requires the i18n interface :)
[19:28] <Chipzz> slangasek: but how many developers would use that anyway?
[19:28] <slangasek> all the ones that I beat with the elder stick
[19:28] <ogra> only the few that actually need to make use of notes :)
[19:28] <Chipzz> and on top, how many translators would bother translating that?
[19:28] <slangasek> all the ones that bubulle beats with the Translation Czar stick
[19:29] <Chipzz> my guess is translatable NEWS.Debian would end up in the back of the translators queue
[19:29] <ogra> or just launchpad if you put it on there :)
[19:29] <Chipzz> s/back of the ... queue/bottom of the pile/
[19:29] <Chipzz> you get the picture :P
[19:30] <slangasek> s/pile/stack/ ;P
[19:30] <Chipzz> anyway, as it stands a lot of HOWTO's on the internet suggest using CA.sh
[19:30] <Chipzz> and don't even mention ssl-cert
[19:32] <Chipzz> but I think I've seen a similar discussion here already (about debconf notes), and I'm not getting my hopes up on it getting fixed :P
[19:32]  * ogra goes to prepare dinner
[19:33] <persia> Chipzz: File a bug anyway, comment on why it's hard, mark it wishlist, and maybe someone will have a good idea.
[19:33] <slangasek> Chipzz: your hopes of getting which part fixed? :)
[19:33] <Chipzz> the ssl-cert visibility thing :)
[19:34] <Chipzz> there was some other issue too, but no-one commented on that yet :P
[19:34] <Chipzz> 20:14 < Chipzz> but this actually brings me to another question: how do the generated certs get used anyway?
[19:34] <Chipzz> 20:15 < Chipzz> since if I manually generate one, it can end up anywhere, for example outside of /etc/ssl, and I don't see how packages using ssl-cert to generate certs  would find the arbitrary name I give it anyway?
[19:42] <slangasek> Chipzz: indeed, if you generate your own cert, you'd either have to put it in place over the snakeoil one, or modify the config of each package to use it
[19:43] <Chipzz> it would need to have the snakeoil name?
[19:44] <slangasek> to be seen without modifying package configs, yes
[19:48] <Chipzz> I'm thinking a symlink to the snakeoil certs, and having packages use the symlinks would be more appropriate?
[19:55] <\sh> slangasek: https://edge.launchpad.net/ubuntu/+source/gail just tell me what went wrong here...looks like I don't get any libgail-common installation candidate on intrepid...regarding lp it's correct ...but thinking about the buildd outputs, it's wrong
[19:56] <\sh> slangasek: looks like the last autosync of gail never build or never been pushed to the archive
[19:57] <\sh> slangasek: forget it ...I hate launchpad
[19:58]  * \sh doesn't get it anymore..
[20:10]  * slangasek eyes the new uninstallables on http://people.ubuntu.com/~ubuntu-archive/testing/intrepid_probs.html
[20:11] <persia> slangasek: Why are matchbox-window-manager and libmatchbox in there?  Those should have been demoted to universe.
[20:11] <slangasek> I don't know, but they weren't there an hour ago
[20:12] <slangasek> hmm, the source was moved to universe, but not the binaries...
[20:12] <persia> Yeah.  libmatchbox was FTBFS for all of intrepid until now, and matchbox-window-manager hadn't been uploaded since gutsy, but FTBFS because of ogre-model.  I requested demotion of both today to get them built.
[20:12] <slangasek> or the binaries were moved to main but not the source, dunno
[20:13] <slangasek> anyway, binaries demoted now, thanks
[20:13] <persia> Probably source demotions without binary demotions.  The archive-admin doing it was new at the job.
[20:13] <slangasek> ah
[20:13] <persia> That ought make them able to be installed again :)
[20:14] <persia> Were those the new ones you were eyeing, or is there something else?
[20:18] <slangasek> StevenK: use -S to grab the binaries too when demoting :)
[20:18] <slangasek> persia: those were the new ones I noticed, yes
[20:19] <persia> Oh good.  I was afraid I broke something else :)
[20:21] <sebner> slangasek: mighty mighty ubuntu and debian release manager. can you tell me an estimated lenny release date? just the month is enough ;)
[20:22] <slangasek> sebner: I'm not a Debian release manager anymore, the Debian RMs have published everything on debian-devel-announce
[20:22] <slangasek> sebner: or you can try to extrapolate based on the graph at http://bugs.debian.org/release-critical/
[20:23] <sebner> slangasek: thx mighty slangasek. I suppose still some months left ^^
[20:24] <Keybuk> sebner: you could fix some and make it faster :p
[20:25] <sebner> Keybuk: and you could include/fix upstart 0.5 to make intrepid faster :P
[20:25] <slangasek> burrn
[20:26] <persia> sebner: That would probably break more than it would fix right now.
[20:26]  * persia would at least have to rewrite a couple things
[20:26] <sebner> persia: we still have alpha status :P :P :P
[20:26] <persia> sebner: Just barely.
[20:26] <Keybuk> sebner: err, why would that make intrepid faster?
[20:27] <slangasek> if we ever want to get /out/ of alpha status, we have to stabilize
[20:27] <Keybuk> Upstart 0.5 missed UVF
[20:27] <Keybuk> and I'm in the same city as slangasek, so I'm not about to break freezes ;)
[20:27] <sebner> You Spoilsport :P
[20:27] <slangasek> hah
[20:27] <Keybuk> at least, not without _asking_ :)
[20:28] <slangasek> oh, is there anything interesting happening downtown tonight, then?
[20:28] <sebner> slangasek: on a tuesday??
[20:28] <Keybuk> there's some event tonight to end the kernel summit and open plumbers
[20:28] <sebner> Keybuk: how does mark want to make jaunty boot faster?
[20:28] <slangasek> on a tuesday the week of the LPC, yes
[20:29] <Keybuk> sebner: push it down a very steep hill ;)
[20:29] <sebner> Keybuk: with a BIGGG kick
[20:30] <Keybuk> I've never understood why people think Upstart makes things faster
[20:30] <Keybuk> it's not supposed to, we've never said it does, and we've denied it every time someone has claimed it would :p
[20:30] <ScottK-laptop> Keybuk: So by the first law of conspiracy thinking it must be true.
[20:31] <sebner> Keybuk: I thought it should replace the old init system (what was old) and makes it faster (what I felt it did)
[20:32] <Keybuk> sebner: ah, you mean the first release of Ubuntu that had Upstart booted much faster than the previous one? :p
[20:32] <Keybuk> I'm afraid that was because we switched from bash to dash as /bin/sh in that release :p
[20:32] <sebner> Keybuk: rofl
[20:32] <sebner> Keybuk: but everybody said that upstart makes things faster xD
[20:32] <Keybuk> (seriously - that was the primary reason we did that, and it worked)
[20:33] <sebner> Keybuk: don't lie. you implemented a secret *boost*-function :P
[20:33] <Keybuk> "Need you now,  buddy!" [TURBO BOOST]
[20:35] <sebner> Keybuk: but it seems to me that every new ubuntu version boots faster. reason? or is it because of the newer kernels?
[20:37] <Keybuk> sebner: it's something we care about
[20:37] <Keybuk> though it's not always true
[20:37] <sebner> Keybuk: yes?
[20:40] <tedg> Keybuk: Do you know why sometimes isn't picking up HAL?  Is it a boot dependency issue?  (mighty annoying one)
[20:40] <tedg> (sorry, for got the "X isn't picking up")
[20:41] <Keybuk> tedg: on Intrepid/
[20:41] <Keybuk> shouldn't be boot order there, dbus is S12, hal is S24 and gdm is S30
[20:42] <soren> ScottK-laptop: Blimey. Building django takes half an hour on my laptop.
[20:42] <sebner> Keybuk: so tell me what your speedup tricks are :P
[20:42] <ScottK-laptop> soren: I'm currently building kdebase on my laptop and considering additional insulation between it and my legs.
[20:43] <tedg> Keybuk: Yeah, on Intrepid :(  It must be a parse time thing then... don't know, definitely happens.  Was worse with the 177 nVidia drivers -- they must be faster :)
[20:43] <Keybuk> sebner: don't waste time doing stuff
[20:43] <tedg> Everyone tells me XML is fast -- I don't see what could slow HAL down.
[20:44] <sebner> Keybuk: no, I mean why does every ubuntu release gets faster (or is this not true)
[20:48] <Keybuk> sebner: as I said, it's not always true
[20:49] <sebner> Keybuk: nice :), and now .. what are the plans to make intrepid faster .. and how?
[20:50] <Keybuk> no plans for intrepid
[20:50] <ScottK-laptop> sebner: Buy an new laptop.
[20:50] <sebner> Keybuk: argh, sry. jaunty
[20:51] <sebner> ScottK-laptop: my laptop is pretty new. but also a quadcore and 4gb ram doesn't make ubuntu boot in 10 seconds :P
[20:52] <Keybuk> sebner: no plans have been drawn up for jaunty yet, that will happen at the UDS in December
[20:53] <sebner> Keybuk: I see .. it's just that I can't say: "Make ubuntu boot in 10 seconds" without knowing how this could be reached
[20:53] <kirkland> seb128: i'm having some intrepid/evolution issues that i'm having trouble describing to Launchpad... after i've read a few messages, and I try to view one in the preview window, the entire message will be greyed out...  double-clicking it to open in a new window helps for a few messages, then those are greyed out too.  restarting evolution allows me to view those messages, but after reading a few more, the same thing happens.
[20:54] <seb128> kirkland: screenshot?
[20:54] <kirkland> seb128: sure, one sec...
[20:54] <Keybuk> sebner: why would you say that?
[20:55] <slangasek> heh, the dbus postinst is comical
[20:55] <slangasek>     # Do not restart dbus on upgrades, only on fresh installations.
[20:55] <slangasek> [...]
[20:55] <slangasek> # Automatically added by dh_installinit
[20:56] <kirkland> seb128: only happens on new, incoming messages... let me way for some more spam to arrive ;-)
[20:56] <sebner> Keybuk: dunno, it's just when *I* have a goal I usually know *if* it's possible and *how* it could be reached
[20:56] <cody-somerville> who said that we'll have Ubuntu booting in 10 seconds? :S
[20:57] <Keybuk> sebner: do you have a goal?
[20:57] <sebner> cody-somerville: that was just an example
[20:57] <sebner> Keybuk: for what?
[20:58] <Keybuk> sebner: sorry, I'm trying to find the context ;)
[20:58] <Keybuk> sebner: whose goal is 10s boot time?
[20:58] <sebner> Keybuk: this was just an example ^^
[20:59] <sebner> Keybuk: Mark wants a faster booting ubuntu and I thought you are the guy to realize it. ^^
[20:59] <Keybuk> sebner: it was a complete surprise to me that Mark talked about boot speed
[20:59] <sebner> Keybuk: so I'm sorry
[21:03] <calc> 10s ubuntu on a regular hd... nice :)
[21:03] <calc> iirc my system is somewhere around 20s already
[21:04] <calc> 10s on a 250MB r/w ssd (doesn't exist yet?) probably is already possible
[21:04] <calc> er 250MB/s
[21:05] <calc> intel's new SSD's are about that for read but not write yet
[21:05] <sebner> calc: with a new intrepid I had 25 secs but now it's dirty and b0rken so I have 35-40 secs xD
[21:05] <calc> sebner: you can retune it, i forgot the command
[21:06] <sebner> calc: nvm. I reinstall it in 1 month
[21:10] <Keybuk> ~25-35s is the lower bound limit of a cold boot right now
[21:10] <Keybuk> faster drives won't speed that up
[21:11] <sebner> Keybuk: windows 7 can boot in 15 secs I heard ^^
[21:12] <Keybuk> sebner: Windows nearly always boots faster
[21:12] <Keybuk> has done for ages
[21:12]  * cody-somerville decides to not make a joke.
[21:12] <sebner> cody-somerville: ^^
[21:12] <sebner> Keybuk: /me never understood *why*
[21:12] <Keybuk> sebner: it's detailed, and technical
[21:13] <cody-somerville> Yea, we had several hours of discussion about it at the last UDS.
[21:13] <Keybuk> but basically, they do most of their stuff after you login ;)
[21:13] <sebner> Keybuk: yes that's an argument
[21:13] <sebner> it's never bad to have a fast boot OS
[21:13] <sebner> fast boot = <20 secs :P
[21:14] <Keybuk> or, put more properly
[21:14] <Keybuk> they do stuff in the right order
[21:14] <Keybuk> and don't waste time doing things because somebody who wears sandals might want to do things wrong
[21:14] <sebner> ^^
[21:15] <sebner> Keybuk: so technically speaking there is a physical barrier with lets say 25secs for a normal kernel/boot? I know a haXX0r can tweak it somehow with selfcompiling and customizing ..
[21:15] <Keybuk> there's a barrier, yes
[21:15] <cody-somerville> and sadly Keybuk is more interested in upstart than boot time :P
[21:15] <Keybuk> that barrier goes away if you compile your own kernel, and build in the drivers you need
[21:16] <Keybuk> cody-somerville: no, keybuk is very interested in boot time also
[21:16] <sebner> Keybuk: so I'm interested if mighty ubuntu folks can accomplish Marks goal O_o
[21:17] <calc> i'm not sure if the 20-25s of my boot is from cold start or not, its what bootchart showed
[21:17] <calc> iirc it supposed to include full boot time
[21:17] <Keybuk> calc: bootchart measures from end-of-grub
[21:17] <ScottK-laptop> slangasek: If you're still around, would you please accept the kdebase in hardy-backports.
[21:17] <calc> Keybuk: oh ok
[21:17] <calc> Keybuk: i thought it grabbed the counter off the cpu to get full time
[21:17] <Keybuk> which is why there's the little white space at the start, which is the kernel starting
[21:17] <Keybuk> calc: we use kernel clock time
[21:17] <calc> ok
[21:18] <Keybuk> (we can't speed up bios :p)
[21:18] <Keybuk> and the grub time is configurable, so uninteresting
[21:18] <calc> true :)
[21:18] <calc> the bios will get taken care of by microsoft
[21:20] <cody-somerville> I remember when the bios was the longest part of the boot process.
[21:20] <azeem> I remember my C64
[21:21] <Keybuk> azeem: your C64 didn't boot
[21:21] <azeem> it said "Booting GEOS"
[21:22] <Adri2000> slangasek: bug #271058
[21:23]  * hardwire pokes soren, respectfully
[21:23] <sebner> bad bios
[21:24] <Keybuk> azeem: it had a ROM in it with all of the operating system code
[21:24] <Keybuk> it just started executing from a prearranged point
[21:25] <Keybuk> and all the hardware was exactly as it was last time
[21:33] <sebner> Keybuk: btw, a quick question. A friend of mine has also ubuntu and before it boots up it takes some seconds. on the screen stands: No raid found, and trying to resume the kernel image. after some seconds it boots normal. any idea? I know here is not a support chan ^^ Btw it's a laptop so I'm wondering about the raid thing
[21:34] <Keybuk> sebner: slow hardware
[21:34] <sebner> Keybuk: it's a brand new Asus G2 ;)
[21:34] <Keybuk> so?
[21:34] <Keybuk> it can still be slow ;)
[21:35] <Keybuk> without further details, it simply sounds like it takes some seconds for his root disk device to be probed, power up, etc.
[21:35] <sebner> Keybuk: if you need futher details just ask. A Asus G2 is a gaming laptop which high end cpu and that stuff. how to fix that?
[21:36] <Keybuk> sebner: means nothing
[21:36] <sebner> Keybuk: in what sense?
[21:36] <Keybuk> in fact, a rule of thumb is the more fancy, singing and dancing a disk controller - the longer it takes to power up
[21:37] <sebner> Keybuk: well, ubuntu but not vista for example :P
[21:37] <Keybuk> you didn't say vista didn't take long
[21:38] <sebner> Keybuk: sry :)
[21:38] <Keybuk> and I suspect, neither did he :p
[21:38] <sebner> Keybuk: actually vista is really booting normal/fast ;)
[21:39] <Keybuk> interesting
[21:40] <sebner> Keybuk: I'm confused by the "No raid found". there isn't a raid thing I suppose and also vista is booting normal
[21:41] <Keybuk> I can't think of any reason either way
[21:41] <Keybuk> it's probably just a message
[21:42] <LaserJock> sebner: I get that message on all of my machines I think
[21:43] <Keybuk> a bit like the resume one
[21:43] <Keybuk> you get a message whether or not you're trying to resume
[21:43] <Keybuk> so I expect you just get that message because there's no raid
[21:43] <LaserJock> exactly
[21:43] <sebner> Keybuk: well, maybe but it take 5-6 seconds until the real boot begins and that's not normal. Also because I'm not seeing it
[21:43] <Keybuk> probably the system reacting to the slow boot, and checking to see whether it's due to failed raid
[21:43] <Keybuk> and because it isn't, saying no raid was found ;)
[21:44] <sebner> Keybuk: isn't the question why it's slowly booting what leads to checking for raid?
[21:44] <Keybuk> sure
[21:44] <Keybuk> it'll be the hard drive taking a while to turn up
[21:45] <Keybuk> it might not be, of course
[21:45] <Keybuk> but that'd be where I'd put my money
[21:45] <sebner> Keybuk: interesting because as I said a gaming pc which should have a pretty good/fast high end hard drice
[21:45] <sebner> *drive
[21:45] <Keybuk> so?
[21:46] <Keybuk> drives are rated on their speed when they're *running*
[21:46] <Keybuk> not speed to spin up
[21:46] <slangasek> ScottK-laptop: accepted
[21:46] <Keybuk> in fact, the faster the drive, the longer it's going to take to spin up!
[21:46] <sebner> Keybuk: but the vista boot question remains
[21:46] <ScottK-laptop> slangasek: Thanks.
[21:47] <Keybuk> sebner: could be driver issue
[21:47] <sebner> Keybuk: so a new kernel *might* fix it?
[21:48] <Keybuk> maybe
[21:48] <Keybuk> worth trying
[21:48] <Keybuk> /var/log/dmesg might be revealing
[21:49] <sebner> Keybuk: I'll tell him. thx :)
[21:50] <Keybuk> scott@ubuntu.com
[21:51] <sebner> Keybuk: hmm?
[21:51] <Keybuk> get him to e-mail it to me there
[21:52] <sebner> Keybuk: well, if you can wait 2 minutes I'll give you a link to a pasteservice ;) (he isn't that good in english so I'm doing this stuff)
[21:52] <Keybuk> sure
[21:55] <sebner> Keybuk: http://paste.ubuntuusers.de/391993/
[21:56] <sebner> Keybuk: could it be that the problem is that he is using 32bit hardy with 4gb ram?
[21:56] <Keybuk> [    0.000000] Detected 2593.872 MHz processor.
[21:56] <Keybuk> [   20.441090] Console: colour VGA+ 80x25
[21:56] <Keybuk> 20s?
[21:56] <Keybuk> sebner: nope, I tend to recommend 32bit
[21:56] <sebner> Keybuk: why?
[21:57] <Keybuk> [   26.870365] usb-storage: probe of 4-1:1.0 failed with error -5
[21:57] <Keybuk> [   26.873351] usb-storage: probe of 4-1:1.1 failed with error -5
[21:57] <sebner> Keybuk: and it's a dualcore
[21:57] <Keybuk> could be a slow USB device?
[21:57] <Keybuk> in fact ...
[21:57] <sebner> Keybuk: hmm I remember that I saw it once. the usb error message
[21:58] <sebner> Keybuk: so what usb error could it be?
[21:59] <Keybuk> in fact
[21:59] <Keybuk> yes, I'd say it's likely that USB device
[21:59] <sebner> Keybuk: hmm, he tells me that the usb error is *after* that thing
[21:59] <Keybuk> if udev is still trying hard to get that probed, it won't mount the disk
[22:00] <Keybuk> sebner: CD-ROM drive maybe
[22:00] <Keybuk> nothing else stands out
[22:00] <sebner> Keybuk: CD-ROM != usb!?!?!
[22:01] <mathiaz> slangasek: landscape-client is currently in NEW. Could you let it there ? It seems that there is an upgrade issue from intrepid.
[22:02] <Keybuk> can you get a "lsusb" output from him and pastebin that?
[22:02] <slangasek> mathiaz: I won't touch it, but I'm not on archive duty today
[22:02] <mathiaz> Riddell: ^^
[22:02] <Keybuk> [   25.876249] usb 5-1: not running at top speed; connect to a high speed hub
[22:02] <sebner> Keybuk: sure
[22:03] <Keybuk> sebner: something storage-related is definitely on USB
[22:03] <sebner> Keybuk: so this might be the boot issue or a seperate issue?
[22:03] <Keybuk> and is definitely slow
[22:03] <Keybuk> I reckon this is your boot issue
[22:03] <Keybuk> it's taking about 4s to probe that device
[22:03] <Keybuk> so that'll show up as a 4s dead period
[22:03] <sebner> Keybuk: also if the error appears later?
[22:03] <Keybuk> error is irrelevant
[22:04] <sebner> kk
[22:04] <sebner> Keybuk: so a controller thing? because nothing external is attached to an USb port
[22:04] <Keybuk> no, there's definitely a device there :p
[22:04] <Keybuk> provide lsusb
[22:04] <Riddell> mathiaz: accepted.  and on my day off too :)
[22:04] <Keybuk> it could be something surprising like a modem
[22:04] <sebner> leyhttp://paste.ubuntuusers.de/391994/
[22:04] <sebner> Keybuk: huwai
[22:05] <Keybuk> that'd be it
[22:05] <sebner> Keybuk: that's the internal modem?
[22:05] <Keybuk> there's a lot of 0000:0000 crap on the bus
[22:05] <Keybuk> sebner: bet it has a usb storage device with the windows driver on it
[22:05] <Keybuk> and I bet the modem isn't even detected by Linux
[22:06] <sebner> Keybuk: so 2 usb issues?
[22:06] <Keybuk> there's some usb issues, yes
[22:06] <Keybuk> the modem will be hardware related, it's probably simply dog slow
[22:06] <sebner> Keybuk: how can a windows driver be on that storage device?
[22:06] <Keybuk> the lack of ids could be hardware, or driver
[22:06] <Keybuk> sebner: all the cool kids are doing it
[22:07] <Keybuk> and by cool kids I mean drug-crazed korean dudes
[22:07] <Keybuk> the USB device appears as a USB storage device with a driver and firmware files on it
[22:07] <Keybuk> once you load the driver, it knows how to kick the device, at which point the USB storage device goes away, and it magically turns into a modem or something
[22:07] <Keybuk> I've seen it for 3G modems
[22:07] <Keybuk> and I've heard of it for TV capture too
[22:09] <ion_> Heh
[22:11] <sebner> Keybuk: so again. a new kernel *might* help? ^^
[22:12] <sebner> Keybuk: so also hardly a new kernel can fix that?
[22:14] <Keybuk> sebner: only may help
[22:15] <Keybuk> the fact it waits for that usb device is a more recent fix to a race condition
[22:15] <Keybuk> which I want to fix a different way
[22:15] <Keybuk> but probably not for jaunty
[22:15] <sebner> Keybuk: hmm he tells me that the internal modem is working without problems ^^
[22:17] <Keybuk> interesting
[22:17] <Keybuk> the HSDPA one?
[22:17] <Keybuk> I didn't think we even vaguely supported those yet
[22:17] <NCommander> Keybuk: HSDPA modems from the computers perspective are the same as EDGE or GSM
[22:18] <Keybuk> NCommander: I know
[22:18] <NCommander> Keybuk: same AT command set
[22:18] <Keybuk> NCommander: the problem isn't their command set
[22:18] <Keybuk> it's their presentation
[22:18] <NCommander> presentation?
[22:19] <sebner> Keybuk: so the problem is a usb thing. maybe not the internal modem but a usb thing?
[22:19] <NCommander> Oh, I see
[22:19] <NCommander> right, some modems show up as a USB device, not a serial port device
[22:19] <Keybuk> sebner: I reckon so
[22:19] <Keybuk> NCommander: it's the fact they don't even show up as modems
[22:19] <NCommander> Keybuk: An anonying problem to say the least. Some don't even support the AT commands :-/
[22:20] <NCommander> And finding specs is (near) impossible
[22:20] <sebner> Keybuk: I see. and you spoke about a fix. dirty fix/good fix?
[22:20] <Keybuk> sebner: the current fix is a good fix, but that introduces a slow down for others
[22:20] <Keybuk> like your friend
[22:21] <Keybuk> there is a better possible fix
[22:21] <sebner> Keybuk: I don't understand. it is not fixed and slow. how can it be fixed and slow?
[22:21] <Keybuk> it is fixed
[22:22] <sebner> Keybuk: in intrepid?
[22:22] <Keybuk> you misunderstand me
[22:22] <Keybuk> there was a bug
[22:22] <Keybuk> which was fixed
[22:22] <Keybuk> but the *fix* means your friend's laptop boots slower
[22:22] <Keybuk> which is arguably better than not booting at all :p
[22:22] <sebner> Keybuk: and tells a error message ... what was the bug? ^^
[22:22] <Keybuk> what error message?
[22:22] <sebner> Keybuk: about that usb thing xD
[22:23] <sebner> Keybuk: so the bugfix fixes the booting but breaks the usb thing? ^^
[22:23] <Keybuk> it doesn't break it
[22:23] <Keybuk> that's broken already
[22:23] <Keybuk> the bugfix just means the boot waits for it to finish
[22:24] <Keybuk> whereas before, it would have carried on even though that usb device wasn't finished
[22:24] <sebner> Keybuk: which is bad?
[22:24] <Keybuk> no, it's not bad
[22:25] <Keybuk> it's simply non-optimal
[22:25] <sebner> Keybuk: faster booting is always nice. nvm on that usb thing
[22:26] <Keybuk> sure
[22:26] <Keybuk> but booting at all is better
[22:27] <sebner> Keybuk: you said: carried on
[22:27] <sebner> Keybuk: or wouldn't it boot without the fix?
[22:27] <Keybuk> someone else's machine doesn't boot without the fix
[22:30] <sebner> Keybuk: I see. so back to the usb issue. it might be fixed with a new kernel that "can" with this usb thing?
[22:31] <Keybuk> I doubt it
[22:31] <Keybuk> but maybe
[22:32] <sebner> Keybuk: sounds bad. so bad asus because of putting windows driver and firmware on it?
[22:33] <Keybuk> nothing to do with asus
[22:34] <sebner> Keybuk: but?
[22:45] <Adri2000> slangasek: ping?
[22:46] <sebner> Keybuk: nvm. He is an idiot xD He isn't believing you. I'm sorry. However. A big thanks from me for taking time :)
[22:49] <cody-somerville> sebner, meh, Keybuk is 50% charm, 50% lies :P
[22:49] <Keybuk> lies?
[22:49] <Keybuk> I don't like
[22:49] <Keybuk> err, like
[22:49] <Keybuk> err, LIE
[22:49] <Keybuk> see, I can't even *spell* lie
[22:50] <Keybuk> I'm frequently mistaken, incorrect or just plain wrong; but that's different
[22:59] <cjwatson> luisbg: your description of your dpkg-divert problem from way back definitely indicates you're using dpkg-divert wrongly, I think (either calling it in the wrong place or with wrong options). dpkg-divert is supposed to work no matter whether you call it before or after the diverted file is installed.
[22:59] <cjwatson> (and indeed if it didn't do so upgrades involving diversions would break really obviously)
[23:01] <NCommander> hey cjwatson
[23:04] <cjwatson> evening
[23:04] <Keybuk> definitely afternoon
[23:04] <NCommander> how goes it?
[23:04] <NCommander> Keybuk: evening here
[23:04] <cjwatson> I have been a bug-fixing machine today. Feeling good about the graphs
[23:05] <cjwatson> </modest
[23:05] <cjwatson> >
[23:05] <NCommander> cjwatson: awesome. I'm a sleep-deprieved machine today :-), although I've been filing bugs on licensing issues I've found and working on improving Nexenta
[23:05] <cody-somerville> It feels like I've gotten very little done today : (
[23:06] <NCommander> cody-somerville: don't worry, that feeling will soon pass
[23:07]  * NCommander hugs an Xubuntu bug :-)
[23:07] <kirkland> cjwatson: thanks for applying the grub-installer fix this AM, that was a dirty little bug ;-)
[23:08] <cjwatson> no problem
[23:08] <cjwatson> the ubiquity graph on http://people.ubuntu.com/~bryce/Plots/ is making me happy at the moment
[23:09] <bryce> cjwatson: almost looks like the US stock market
[23:09] <cjwatson> ha
[23:10] <james_w> cjwatson: hi, I looked at luisbg's source package, and it was calling dpkg-divert the way that is suggested in debian policy, modulo argument ordering, which the code suggested would make no difference.
[23:10] <cjwatson> then I'm pretty confused - not sure I have time to look at it right now though
[23:11] <james_w> however, if dpkg-divert is supposed to work in that case then I agree there must be something wrong
[23:11] <cjwatson> if memory serves you do dpkg-divert --blahdeblah --rename --divert foo.orig foo
[23:11] <cjwatson> in the preinst
[23:12] <james_w> yeah, that's what was in policy
[23:13] <NCommander> Oh that's lovely
[23:13]  * NCommander hates his college
[23:24] <TheMuso> asac: Its on my radar. Theres a couple of different solutions I'm looking into.
[23:33] <TheMuso> NCommander: You don't care about gnucash FTBFSing on hppa do you?
[23:33] <NCommander> TheMuso: thats a transiant failure, its a misidentified depw-ait
[23:34] <NCommander> TheMuso: its not like I can do much on hppa failures; I don't have access to the hardware
[23:34] <TheMuso> NCommander: oh ok you know about it then. I'll leve it up to you to sort out.
[23:34] <NCommander> TheMuso: It needs to be retried once the build-dep is in place
[23:34] <TheMuso> NCommander: ok.
[23:41] <cody-somerville> NCommander, have you become the FTBFS-go-to-guy? :)
[23:41] <NCommander> cody-somerville: it seems so
[23:43] <NCommander> cody-somerville: where is the FTBFS?
[23:43] <TheMuso> cody-somerville: In this case, I uploaded gnucash for NCommander, so was just checking he knew of the FTBFs.
[23:44] <NCommander> There isn't much I can do with it since hppa has fallen behind again :-/
[23:44] <cjwatson> glib2.0 FTBFS on hppa; fixing that would help
[23:45] <asac> TheMuso: ok. you think we can get an initial solution for broader testing before beta?
[23:45] <NCommander> cjwatson: know where I can find a public HPPA box running Ubuntu? (or debian with a pbuilder ubuntu chroot_
[23:45] <cjwatson> not I, lamont might
[23:45] <NCommander> lamont said he can't give shell access to his machines ATM
[23:45] <TheMuso> asac: I hope so. Thats what I'm planning on.
[23:47] <cjwatson> I don't know of an hppa porter box for Canonical developers, even
[23:48] <cjwatson> though the list I'm looking at has been known to be behind
[23:48] <TheMuso> Frankly, i'm not bothered, but since it was NCommander's upload, I thought he migh care. :)
[23:51] <cody-somerville> To be honest, I'm really broken up over it. : (
[23:51] <TheMuso> heh
[23:52] <NCommander> cjwatson: I'm suprised hppa is a port due to the distinct lack of hardware. We've been trying to fix an FTBFS in KDE for HPPA, and we've currently had three uploads, and three misses
[23:56] <NCommander> cjwatson: Well, I have a pending developer account request with HP for remote access to one of their hppa boxs
[23:57] <cody-somerville> I think I already have one myselgf
[23:57] <NCommander> cody-somerville: what, hppa?
[23:57] <cody-somerville> I think
[23:57] <NCommander> "think"?
[23:58] <NCommander> cody-somerville: can I possibly buy it off you? (I'll drive to come and get it ;-))
[23:58] <cody-somerville> No, I think I have a shell account from HP
[23:58] <NCommander> Oh :-P
[23:58] <cody-somerville> but I may know someone with an actually hppa box.
[23:59] <cody-somerville> erm
[23:59] <cody-somerville> *actual
[23:59] <NCommander> Well, it hopefully if this request is approved, I'll get both IA64 and HPPA accounts
[23:59] <NCommander> So I might be able to help fix fun FTBFS on all architectures