[00:11] <GanonKiller> can anybody here help me?
[00:14] <TheMuso> GanonKiller: Just ask your questino. If people are around and can assist, they will.
[00:14] <TheMuso> question
[00:15] <GanonKiller> well pulseaudio keeps crashing
[00:16] <GanonKiller> i am trying to fix it
[00:16] <GanonKiller> i removed pulseaudio before & it killed gnome
[00:16] <TheMuso> How does it keep crashing? have you filed a bug?
[00:16] <GanonKiller> yes i have filed a bug report
[00:16] <TheMuso> Do you have the bug number handy?
[00:17] <GanonKiller> how do i find that out?
[00:19] <TheMuso> GanonKiller: If you know your launchpad username, then go to http://launchpad.net/~username/+subscribedbugs and you should find it there, unless you decided to unsubscribe yourself from the bug once you filed it. :)
[00:21] <GanonKiller> i subscribed ... i just dont remember my info
[00:23] <TheMuso> GanonKiller: Check your email, you might have some email regarding the bug, which can tell you what bug number it is.
[00:24] <GanonKiller> ok... well i sent the bug report via terminal command
[00:25] <TheMuso> Ok, let me see fi i can find it...
[00:26] <bencc> why ubuntu doesn't download and install packages in parallel?
[00:27] <GanonKiller> well.. i think i need to change my sound profile
[00:28] <TheMuso> bencc: If you are using more than one mirror, I believe apt can download in parallel, as to why it doesn't download in parallel normally, I am ont entirely sure. As for installing, packages have to be installed in a specific order, to make sure dependencies are correctly resolved, as the install process of one package may rely on a previously installed package to complete.
[00:29] <bencc> TheMuso: the download order could be optimized so apt can install undependent packages while it downloads others
[00:29] <TheMuso> bencc: For parallel download I agree, but not for installing.
[00:29] <TheMuso> It needs all packages to be present to be able to work out the installation order.
[00:30] <bencc> TheMuso: I don't think so
[00:30] <bencc> TheMuso: if you want to upgrade apache and mysql at the same time
[00:30] <TheMuso> bencc: Others know more about this than I do, but I am pretty sure that this is just not possible.
[00:30] <bencc> it's possible that they don't share the same package so you can start upgrading mysql while downloading apache
[00:32] <bencc> TheMuso: I'm 99.99% positive it's posible
[00:32] <bencc> and I'm 100% positive it'll be cool
[00:32] <TheMuso> bencc: You might want to read up on apt and dpkg then, to find out how they currently do things.
[00:33] <bencc> too complicated...
[00:34] <RAOF> :)
[00:34] <bencc> the fact that it's complicated doesn't mean my idea isn't simple or can't be done
[00:35] <bencc> trust me, it's the best thing since slice bread
[00:35] <RAOF> I would have thought the implementation of your idea being complicated *did* mean that the idea wasn't simple, in practice :)
[00:36] <RAOF> You'd need to do some deep dpkg diving; there's a bunch of global state in the form of package databases that you'd need to be sure won't be corrupted by multiple accesses.
[00:36] <bencc> RAOF: make things as simple as possible but not simpler than that
[00:36] <bencc> RAOF: now you are complicating stuff
[00:36] <bencc> I don't need multiple accesses
[00:37] <RAOF> Yes, you do.
[00:37] <bencc> no, I don't
[00:37] <GanonKiller> which is better... tunapie2 or rhythmbox?
[00:37] <bencc> you are just arguing for the sake of argument
[00:37] <TheMuso> GanonKiller: Ok I can't find any bugs that were recently filed against pulseaudio that have you're name, Dewey Van (Taken from whois on IRC). So unless you have a bug number, I am unable to help you further at this point.
[00:37] <bencc> probably tunapie2, because the 2
[00:37] <GanonKiller> ok
[00:37] <bencc> just kidding
[00:37] <RAOF> You're installing two packages at once; that means you need to be able to access the package database (and debconf database, and such) from two separate installs.
[00:38] <bencc> RAOF: I'm not
[00:38] <GanonKiller> i will take off rhythmbox
[00:38] <RAOF> You're not installing two packages at once?
[00:38] <RAOF> I thought that was your idea?
[00:38] <bencc> RAOF: no. I'm downloading and installing at the same time
[00:38] <bencc> it's different
[00:39] <bencc> installing two packages at once will be stupid
[00:39] <bencc> (I'm not)
[00:39] <RAOF> That would indeed be easier; it's probably not *that* difficult.
[00:40] <bencc> right
[00:40] <bencc> and it's very very cool
[00:40] <RAOF> Why don't you find out, by writing the patch? :)
[00:40] <bencc> maybe I will
[00:40] <bencc> do you want to help me with that
[00:40] <bencc> to share the glory
[00:40] <bencc> ?
[00:41] <GanonKiller> ok.. i installing the other pulseaudio devices & removing rhythmbox
[00:41] <RAOF> bencc: No, sorry; I've got other projects on my plate.
[00:41] <bencc> RAOF: like what?
[00:41] <bencc> what can possibly be more important than that?
[00:42] <TheMuso> ...like making your graphics work properly?
[00:42] <taggart> bencc: it's usually a good idea to have all the debs sitting on your disk before starting to install, in case something goes wrong that fucks up the network or something
[00:42] <RAOF> Any number of things?  C# bindings for libpulse, hacking on Do, gobject-introspection bindings for C#...
[00:42] <taggart> the odds of that are low, but I've definitely had situations where I was glad they were already there
[00:42] <GanonKiller> is there to invert the 3d desktop cube?
[00:43] <TheMuso> taggart: Please be careful of your language in here, and abide by the Ubuntu code of conduct.
[00:43] <taggart> actually for really big upgrades I do "apt-get -d upgrade" first and get all the packages
[00:43] <taggart> TheMuso: bite me
[00:43] <bencc> RAOF: what is Do?
[00:44] <RAOF> http://do.davebsd.com/
[00:44] <RAOF> bencc: Your idea sounds cool, but it's somewhat of a micro-optimisation.
[00:44] <GanonKiller> hmmmm the alsa plugin is having issues... the volume is moving on its own
[00:44] <bencc> RAOF: ok
[00:45] <bencc> Do looks cool
[00:46] <ion> themuso: The CoC doesn’t explicitly say anything about which words are good and which are bad. That’s very subjective.
[00:47] <TheMuso> ion: Ok.
[00:47] <bencc> I'm sure they don't want us to say fuck
[00:47] <bencc> oops :)
[00:47] <RAOF> I think it's a pretty safe call that the word in question falls under ‘bad’.
[00:48] <ion> Okay, i admit it was disrespectful towards the package that broke network. taggart: Please apologize to the package.
[00:50] <persia> No.  There are no rules for that in the CoC, that's about respect and raising disputes quickly and avoiding personal attacks and stuff.
[00:50] <persia> Not using certain words is about our IRC guidelines.
[00:50] <persia> !ohmy
[00:53] <TheMuso> GanonKiller: When you say your alsa volume is moving on its own, what program are you looking at and using when it does this?
[00:58] <GanonKiller> i forgot where to change the sound type?
[00:58] <GanonKiller> oh lol... pulse audio volume control
[01:03] <GanonKiller> dang sound
[01:04] <GanonKiller> i did install the vlc-plugin-pulse
[01:04] <GanonKiller> the alsa-playback volume keeps moving on its own
[01:04] <BUGa_29plus1> one of this days I'll get why PM tools work fine, and system suspend/hibernate doesn't
[01:05] <TheMuso> GanonKiller: When you are using vlc?
[01:05] <GanonKiller> i am using VLC right now with tunapie2
[01:07] <GanonKiller> how do i restart the sound server'/
[01:07] <TheMuso> GanonKiller: Why do you want to restart it?
[01:08] <GanonKiller> it killed
[01:09] <TheMuso> right, as you previously stated.
[01:09] <GanonKiller> sudo: /etc/init.d/alsa-utils: command not found
[01:10] <GanonKiller> i cant restart
[01:11] <TheMuso> GanonKiller: Please reproduce the problem following this procedure: https://wiki.ubuntu.com/PulseAudio/Log and !pastebin it somewhere.
[01:11] <TheMuso> !pastebin
[01:19] <GanonKiller> heh... now my sound works
[01:22] <GanonKiller> it turns the sound controller was controlling the app
[01:22] <GanonKiller> *turns out
[01:22] <GanonKiller> the volume was all the way down
[01:23] <GanonKiller> i also had to change the output
[01:24] <GanonKiller> thanks for your help muso
[02:32] <dasKreech> Hallo. Who would I speak to about a dead official Ubuntu server?
[02:33] <persia> Which one?
[02:33] <dasKreech> speglar.simnet.is
[02:36] <dasKreech> Let me call the person with the issue
[02:36] <dasKreech> persia: This is the person querying the server
[02:36] <dasKreech> You can get more information from them
[02:36] <Guest64905> This is Reykjavik,,,, do you read me?  over.........
[02:36] <persia> You could send a note to mirrors@ubuntu.com.  Someone might be active in #ubuntu-mirrors.  The admins at Siminn hf are the folks that really need to sort it.
[02:36] <dasKreech> :-)
[02:37]  * dasKreech leaves persia and Guest64905 :)
[02:37] <Guest64905> Yea thank you
[02:38] <Guest64905> Simnet.is is a big ISP  its dificault to talk to them... the other server is in the uneversety
[02:38] <Guest64905> ubuntu.lhi.is
[02:38] <persia> Guest64905, So, it looks like the administrator for that mirror has no public address on LP.  You could try sending a note to mirrors@ubuntu.com reporting the mirror outage.  You might find someone active in #ubuntu-mirrors (but I suspect not the Reykyavik admins at this hout)
[02:38] <Guest64905> Thank you I try
[02:38] <persia> You could try using https://launchpad.net/ubuntu/+mirror/speglar.simnet.is-archive has a base set of information to contact the contact person, but without a public address, I presume they don't prefer to be contacted that way.
[02:39] <persia> Good luck!
[02:41] <RAOF> Oh, accursed heisenbug.
[02:42] <Guest64905> I am sending mirrors@ubuntu.com a message.... the ubuntu.lhi.is  server is in the Icelandic part of this huge networ
[02:42] <Guest64905> http://www.nordu.net/ndnweb/nordunet___canarie_announces_icelink.html
[02:42] <RAOF> What are my options for hunting down a memory corruption bug?  Bug #651294 seems to be caused by something outside the resource code scribbling random(?) values over the memory held by the resource pointer, leading to badness when trying to free the resource.
[03:25] <persia> RAOF, Well, if you happen to have a copy of the entire memory space, you can hope to find something meaningful in the otherwise random scribbling.
[03:25] <RAOF> Or, I guess, set up some gdb watches on that memory address, and just eat the abysmal performance.
[03:25] <persia> If you can reproduce, you can run large chunks of stuff through memory profilers (e.g. valgrind) at fairly slow speeds.
[03:26] <persia> gdb is less abysmal than valgrind :)
[03:26] <persia> But if you know the address that's being scribbled, it's not quite random ...
[03:26] <RAOF> Oh, no.
[03:27] <RAOF> The address is fixed; what's beeing written there seems random.
[03:27] <RAOF> Or, rather, I know the address of the resource struct, and somewhere between being added and being freed something is scribbling on it.
[03:27] <persia> In that case, you don't even need gdb: you could just poll that address repeatedly and spit out some output when it changes somewhere.
[03:28] <persia> (still performance-bad, but less bad than all of gdb)
[03:29] <RAOF> I might try the gdb watch approach; as I understand it, that _should_ get me a backtrace of what's actually doing the scribbling.
[03:29] <RAOF> Sadly, it also means logging out of X will take in excess of 12 hours.
[03:30] <persia> I doubt it.
[03:30] <persia> Unless you are also running the scribbling bit inside gdb.
[03:30] <persia> Depends whether your issue is localised to a single process or not.
[03:31] <RAOF> I presume it is, but I guess that it's not guaranteed to be the case
[03:32] <RAOF> Barring a kernel bug, though, a second process couldn't scribble over X's memory, right?
[03:32] <RAOF> I guess it decided that poking /dev/mem or whatever it is is a fun way of communicating.
[03:33] <TheMuso> Wow. New binutils in natty is calling for more bits to be explicitly linked... At least thats what I am currently looking at in a test pulseaudio build...
[03:38] <persia> RAOF, char *x = 32489716; x='Q';
[03:38] <RAOF> SIGSEGV?
[03:39] <persia> Yeah, well, if it was that clean, we wouldn't have to worry so much about PIC and PIE :)
[03:39]  * persia isn't going to show something that really works here, and points at exploit documentation for those really interested
[03:40] <persia> But if you can reproduce it reliably, there's a good chance that it is one process.
[03:40] <RAOF> Yeah.  It's going to be something in the X server.
[03:40] <RAOF> Or, I guess it could be kwin doing something totally rediculous.
[03:44] <persia> Could try trapping each with GDB over a multi-day period (one try each), and see if you can catch it.
[04:19] <GanonKiller> i am having trouble opening a rar file
[04:19] <GanonKiller> i have have 7zip installed & cant find the prog
[04:19] <TheMuso> GanonKiller: #ubuntu for support, you were helped earlier because you had a crash, but this is a general support query.
[04:20] <GanonKiller> sorry
[04:20] <TheMuso> This channel is for development.
[04:20] <TheMuso> No proble,/.
[04:20] <TheMuso> problem
[05:13] <robert_ancell> Is natty open for uploading?  I can't find any information either way...
[05:13] <nigelb> robert_ancell: /topic in -motu says it is
[05:14] <robert_ancell> nigelb, thansk
[05:15] <nigelb> np
[05:19] <persia> robert_ancell, The formal answer to that question is always available from LP: e.g. https://launchpad.net/ubuntu/natty
[05:21] <robert_ancell> persia, yeah, I've been looking at that but would it show there if it was in a freeze?
[05:23] <persia> robert_ancell, "Status" was "Pre-release Freeze" about 6 hours ago.
[05:23] <persia> https://launchpad.net/ubuntu/maverick may be an interesting comparison
[05:24] <robert_ancell> persia, ah, thanks
[06:26] <bilalakhtar> Why not change this channel topic to reflect the fact that natty is open?
[06:26] <persia> Just nobody did it.
[06:26] <bilalakhtar> persia: I am doing it now
[06:27] <persia> Oh, heh, me too :)
[06:27] <bilalakhtar> :D
[06:28] <mneptok> persia: günaydin, effendi. nasilsin?
[06:30] <persia> Decidedly unurgaric :p
[07:35] <pitti> Good morning
[07:35] <ion> that.
[07:35] <pitti> YokoZar: what was the bug ref # again?
[07:36] <YokoZar> pitti: 6222220
[07:36] <YokoZar> err 622220
[07:38] <persia> RAOF, Would you have an interest in https://blueprints.launchpad.net/ubuntu/+spec/hardware-arm-n-irregular-input-device-handler ?  Might be udev hinting, might be fiddling with gizmod or similar.
[07:41] <RAOF> Hm.  Looks interesting, but I've no experience with arm hardware.
[07:41] <persia> From the perspective of knobs and buttons, it's not that different from anything else.
[07:42] <persia> Just /dev/input/event entries.
[07:42] <pitti> YokoZar: uploaded
[07:42] <persia> Adding "arm" is mostly intended to explain why someone should consider it critically important :)
[07:42] <YokoZar> pitti: thanks
[07:54] <pitti> hooray, natty thawed?
[07:54]  * pitti yays at the wave of "accepted" mails
[07:57] <persia> Almost 9 hours ago :)
[07:57] <dholbach> good morning! :)
[07:59] <bilalakhtar> dholbach: Good morning!
[07:59] <dholbach> hey bilalakhtar
[08:54] <smb> pitti, There is a ti-omap4 proposed kernel for maverick waiting to be accepted which ogra has been winching about all along yesterday. Could you do him the favor and accept it? ;-)
[08:54] <pitti> right, it's been 15 hours since I processed SRU, high time :)
[08:54]  * pitti looks
[08:55]  * pitti tries to ignore the gazillion new .gitignore files
[08:56] <pitti> smb: hm, why does this have several new source files, like arch/arm/mach-omap2/powerdomain.c ?
[08:56] <pitti> or arch/powerpc/include/asm/immap_86xx.h (must be really important for omap..)
[08:56] <smb> pitti, could this be a result of me doing a orig.tar.gz the first time?
[08:57] <pitti> ah, perhaps
[08:57] <pitti> sound/soc/omap/mcpdm.[hc] are also new
[08:58] <pitti> so, this needs some more testing then
[08:58] <smb> pitti, Oh, actually they did some sound stuff
[08:58] <smb> but as it is arm I have no idea what they do anyways
[08:58] <pitti> not in the changelog
[08:59] <pitti> anyway, sending builddwards
[08:59] <smb> pitti, Thanks, they seemed quite keen to have it till Monday for some reason
[09:00] <smb> and as it is a topic branch and this kernel is only used on some few board, regression potential is low
[09:15] <persia> smb, pitti Regression potential is effectively zero, as nothing that kernel supports is currently available retail.
[09:15] <pitti> right, that's why I'm not too concerned about it; I just wondered about the diff
[09:17] <persia> From the bug descriptions, sound/soc/omap/mcpdm.[hc] are required to make sound work at all.
[09:22] <ogra> pitti, that kernel change goes together with the alsa-utils upload
[09:22] <pitti> ah
[09:22] <ogra> (same bug)
[09:33] <lool> doko: I have a build of newlib for ppc in a native PPA; do you want the build log over email to proof-check it?
[09:57] <ogra> mvo, is sw-center set to "always on top" ? (see bug 661003)
[10:02] <mvo> ogra: no, but it sounds like debconf should be
[10:02] <mvo> ogra: its probably focus stealing prevention
[10:04] <ogra> mvo, hmm, k
[10:05] <mvo> ogra: and yes, not cool ./
[10:06] <ogra> well, minor usability glitch
[10:06] <ogra> would be worse if it was not shown at all :)
[10:06] <persia> for some values of minor.
[10:06] <mvo> true that
[10:06] <mvo> (the not shown at all)
[10:14] <cjwatson> TheMuso: your linking question: see http://wiki.debian.org/ToolChain/DSOLinking
[10:14] <cjwatson> doko: ^- you might want to include a note about that in your mail about natty opening
[10:23] <doko> lool: fine it it does build
[10:23] <doko> cjwatson: yes, have to update it first
[10:27] <poolie> hi
[10:28] <poolie> can someone _please_ help us get SRUs for https://bugs.edge.launchpad.net/bzr/+bug/636930 ?
[10:45] <cjwatson> poolie: sponsoring; will still need approval from another ubuntu-sru team member
[10:45] <cjwatson> but this will move it forward at least
[10:45] <poolie> thanks
[10:45] <pitti> I'll review it
[10:45] <poolie> it doesn't have to be today in particular
[10:45] <poolie> i just don't want it to stall forever
[10:45] <pitti> I'm doing SRU review every day now anyway, due to the maverick rush
[10:46] <poolie> pitti, how would i find out or establish a regular contact for this?
[10:46] <\sh> mom is already running?
[10:47] <cjwatson> \sh: yes
[10:47] <pitti> poolie: hm, not sure; we should perhaps discuss on u-devel@, to find a permanent sponsor
[10:47] <\sh> cjwatson: thx :)
[10:48] <cjwatson> might be better for maxb to seek per-package upload rights, if he's going to be a regular uploader
[10:48] <poolie> there are ubuntu devels who upload it, but nobody in the canonical platform team, and i think nobody who will regularly SRU
[10:48] <poolie> i'm going to work on getting per-package upload rights myself
[10:49] <cjwatson> doesn't have to be the canonical platform team
[10:49] <pitti> seems so far we mostly got them through Debian autosyncs?
[10:50] <poolie> right, which is good, i think
[10:50] <cjwatson> definitely
[10:50] <poolie> that works pretty well, but the main snag is SRUs
[10:50] <cjwatson> (in fact, I spy 2.3.0~beta2-1 in the pending natty sync queue)
[10:50] <poolie> we really want to do bugfix releases that fit well with SRUs, but i always seem to have to pester people to get them sponsored
[10:50] <poolie> so i think we must be doing something wrong
[10:51] <alkisg> Which Ubuntu package inserts "%admin ALL=(ALL) ALL" to /etc/sudoers? That line is not present in Ubuntu fat chroots...
[10:51] <maxb> What do I need to do to be eligible for per-package upload rights? I've been getting bits and pieces sponsored here and there for a while, but never quite in sufficient volume to consider myself MOTU-eligible.
[10:51] <cjwatson> any time we find ourselves assembling a permanent sponsorship arrangement, we should probably be reaching for per-package upload instead
[10:51] <lool> doko: Thanks; uploading
[10:51] <cjwatson> alkisg: the installer - specifically user-setup
[10:51] <alkisg> Thank you cjwatson
[10:51] <maxb> Not, of course, that MOTU would help for bzr
[10:51] <cjwatson> maxb: https://wiki.ubuntu.com/UbuntuDevelopers#PerPackage
[10:52] <pitti> maxb: your sponsor needs to vouch that you have sufficient experience with the package you want to upload yourself
[10:52] <pitti> maxb: that's the main bit, and then some paperwork
[10:52] <cjwatson> poolie would surely be a shoe-in for bzr upload rights, and I expect maxb wouldn't have much difficulty either
[10:52] <cjwatson> (shoo-in?)
[10:52] <persia> Sponsoring of SRUs has always lagged a lot: many folk don't seem to focus on that.
[10:56] <ari-tczew> now should be enough subscribe ubuntu-sponsors. devs can review bug and patch. if it's acceptable, can upload
[10:56] <ari-tczew> acceptable => test case exists in bug, patch is correct with policy and built fine
[10:56] <poolie> i know, but it basically never seems to happen without coming in here and asking several times
[10:56] <poolie> i know you're not lazy, so it suggests something is disconnected
[10:57] <persia> poolie, I suspect it's a perception that bzr is cared for by someone, and so everyone leaves it to whoever cares for bzr.
[10:57] <persia> Except, since nobody actually cares for bzr, this fails.
[10:57] <persia> per-package upload rights is definitely the right fix in this case.
[10:58] <poolie> i don't think regular uploads are a problem, it's only or mostly SRUs
[10:58] <poolie> and aiui per-package uploads won't help with that? or will they
[10:58] <persia> I believe they help from the date that you get per-package.
[10:58] <pitti> poolie: PPU works just as well for SRUs
[10:59] <persia> So it may not help with maverick, but it will help with natty, ... once they release.
[10:59] <persia> pitti, Does it?  All the way back?
[10:59] <pitti> persia: it's another step in edit-acl.py, but yes
[10:59]  * persia was under the impression that packagesets were per-release
[10:59] <poolie> ok, brilliant, we should definitely do that then
[10:59] <pitti> persia: they are
[10:59] <persia> pitti, Ah, so it *can* be done that way :)
[10:59] <pitti> persia: but we can edit them per-release
[10:59] <cjwatson> PPU isn't packagesets though
[10:59] <cjwatson> give me an example of somebody with PPU?
[10:59] <persia> cjwatson, It isn't?
[10:59] <cjwatson> preferably LP username
[11:00] <persia> cyphermox
[11:00] <pitti> tkamppeter?
[11:00] <persia> Except we fail
[11:01] <pitti> till-kamppeter is the LP username
[11:01] <persia> ~till-kamppeter
[11:01] <cjwatson> sorry, I should say, PPU isn't necessarily packagesets.  it's perfectly possible to grant people direct upload access to a single package without creating a set for it
[11:01] <cjwatson> archive permissions can be attached to components, packagesets, or packages
[11:02] <cjwatson> cyphermox => ~mathieu-tl, and he has upload rights via team membership for a package set; package set access does have to be release-specific
[11:02] <poolie> so with PPU, maxb or I could upload to -proposed, and then someone from ~ubuntu-sru would check it, and copy to -updates?
[11:02] <cjwatson> (though we can always go back and extend things for older releases)
[11:02] <persia> cjwatson, Did all of cyphermox's permissions get moved to packageset?  He applied for PPU something like three times.
[11:02] <cjwatson> ~till-kamppeter has PPU access to several source packages, and that access isn't release-specific
[11:02] <cjwatson> persia: yes, I checked
[11:02] <persia> Ah, OK.
[11:03] <cjwatson> poolie: right
[11:03] <cjwatson> (modulo LP bugs; not many people use per-package upload permissions for SRUs, so you may be trailblazing)
[11:03] <persia> poolie, Note that this is the same procedure every Ubuntu developer needs to follow (even SRU members, who ask their peers to do the second part)
[11:03] <cjwatson> right.  the SRU queue is better tended than the sponsors queue, I think, largely thanks to pitti being a hero
[11:04] <maxb> poolie: You missed a step, we could upload to -proposed, it would hit the queue, ~ubuntu-sru would accept it, it would need verification, and *then* it would be copied to -updates
[11:05] <persia> maxb, Extra points for process knowledge !
[11:05] <poolie> ok, i was oversimplifying a bit; but the big difference is that as cjwatson said it will get into a queue that's more actively polled
[11:05] <poolie> per https://wiki.ubuntu.com/ArchiveAdministration#Stable release updates
[11:05] <cjwatson> and a step forward in the process
[11:05] <poolie> so if you two will be so kind as to put this one into proposed, we'll see about getting PPU for next time
[11:06] <poolie> and 2.2.2 should also have clean test suites both in package build and after installation
[11:07] <persia> poolie, And just to be clear, PPU is a one-time hurdle.  We'll expect you to act as an Ubuntu Developer in full faith once it's granted, but you won't have to revisit the process repeatedly (except potentially iterating per-person)
[11:08] <poolie> sure
[11:08] <poolie> i have fixed a few bugs in other packages recently
[11:08] <poolie> well, for years, but more so recently
[11:08] <poolie> so, perhaps i should go for something more general afterwards
[11:11] <persia> Best to apply for things as you find that folks sponsoring you do so confusedly expecting you to have done it already :)  You'll likely end up with larger and larger portions of the archive over time.
[11:12]  * cjwatson flushes the initial natty sync queue
[11:12] <poolie> :)
[11:12] <cjwatson> I only had to beat sync-source.py with a stick once.  Do you think Debian is frozen?
[11:13] <lool> I don't understand why https://launchpad.net/ubuntu/+source/newlib/1.18.0-6ubuntu1 doesn't have any armel and amd64 build records
[11:14] <lool> it did build on amd64 and armel in a non-virtual PPA
[11:14] <cjwatson> it's in Packages-arch-specific
[11:14] <cjwatson> %newlib: i386 powerpc ppc64
[11:14] <bilalakhtar> doko_: there? What about the change you mailed about?
[11:15] <lool> cjwatson: Oh ok; this is new then, as ithad built in maverick
[11:15] <cjwatson> does it build on other architectures in Debian?  if so, the best fix would be to file a bug on buildd.debian.org about that
[11:15] <cjwatson> lool: see also http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=571256
[11:16] <lool> Yup, just found it in the Pas git log
[11:16] <cjwatson> hm, the bug appears to be "it builds but I didn't think it was useful"
[11:16] <cjwatson> maybe talk to Arthur about that?
[11:16] <lool> why no explanation?
[11:16] <cjwatson> read down to the end of the bug
[11:16] <cjwatson> (that's all I know)
[11:16]  * lool searches for RM bug
[11:17] <lool> Debian #597234
[11:17] <lool> not anything more
[11:19] <lool> I poked the Pas bug
[11:19] <cjwatson> hence my suggestion to talk to Arthur :)
[11:20] <\sh> guys, when I get such a build error like http://paste.ubuntu.com/513753/ <- what needs to be done for natty? I saw some uploads by Doko with a similar changelog entry...is there any doc on how to fix this?
[11:21] <ari-tczew> someone ping me?
[11:21] <bilalakhtar> ari-tczew: 13:19 < cjwatson> hence my suggestion to talk to Arthur :)
[11:23] <TheMuso> cjwatson: thanks
[11:24] <persia> \sh, maybe http://wiki.debian.org/ToolChain/DSOLinking ?
[11:24] <cjwatson> ari-tczew: I didn't mean you (I'd have spelled your name right :-) ); I was referring to the Debian bug above
[11:25] <gusnan> My GTK program get the following: http://paste.ubuntu.com/513757/ during build - does someone know anything about this?
[11:25] <bilalakhtar> gusnan: install the libgdk-pixbuf packages
[11:25] <bilalakhtar> gusnan: but for such questions, go to #ubuntu-app-devel
[11:25] <bilalakhtar> not here
[11:25] <\sh> persia: thx a lot friend :)
[11:26]  * persia is just an advanced backscroll filter sometimes
[11:26] <ari-tczew> cjwatson: ok then :)
[11:26] <\sh> persia: lol
[11:26] <gusnan> bilalakhtar, thanks.
[11:52] <tkamppeter> pitti, hi
[11:56] <pitti> hey tkamppeter, how are you?
[11:57] <tkamppeter> fine, have you asked for me?
[11:57] <pitti> no, I didn't
[11:57] <persia> tkamppeter, We were using you as an example of a per-package uploader.  Sorry for the noise.
[11:57] <pitti> ah, right
[11:57] <tkamppeter> pitti, what about bug 485383?
[11:59] <tkamppeter> pitti, as you ahve all kinds of BZR repos for that, can you prepare a CUPS SRU only adding a cups-ppdc dependency to cups?
[11:59] <pitti> tkamppeter: feel free to push to bzr+ssh://bazaar.launchpad.net/~ubuntu-printing/cups/maverick/
[11:59] <pitti> tkamppeter: (sorry, we have a deadline today, no time for Ubuntu)
[12:00] <tkamppeter> pitti, for Debian?
[12:00] <pitti> neither, I'm afraid :)
[12:08] <tkamppeter> pitti, pushed it up now, rev 910.
[12:18] <tkamppeter> pitti, SRU is uploaded to the queue now.
[12:31] <ari-tczew> doko: ping
[12:41] <sladen> mvo: would it be feasible to tweak update-manager to check for machine >= i586 and refuse to attempt the upgrade if it is noticed?
[12:42] <sladen> mvo: at the moment function 10.04 machines get hosed on upgrade to 10.10 once libc6 gets installed and starts generating illegal instruction traps
[12:43] <mvo> sladen: *gar*
[12:43] <mvo> sladen: sure, definitely
[12:43] <mvo> sladen: do you have more info what to check for in cpuinfo ? or a bug reference?
[12:43] <mvo> sladen: not the kind of message I was hoping for on a friday afternoon ;)
[12:44] <Ian_Corne> :)
[12:44] <sladen> mvo: https://bugs.launchpad.net/ubuntu/+source/eglibc/+bug/587186
[12:44] <mvo> sladen: thanks!
[12:49] <mvo> sladen: so it looks like checking the flags section for cmov should do the trick?
[12:50] <persia> Does 10.10 run on the VIA C7M chips?  I seem to remember they weren't quite i686 but had cmov.
[12:50] <persia> (mind you, I'm remembering from intrepid, so I might be incorrect: C7M was the chip in early EeePCs)
[12:51] <mvo> persia: I don't know :) any help on this is welcome how to reliable detect it
[12:51] <mvo> persia: so either cpufamilty 5 -> fail or cpufamily 6+ and not cmov -> fail?
[12:51] <persia> Anyone have a first-generation EeePC or similar device with C7M?
[12:51] <mvo> (5 or less)
[12:52] <persia> mvo, I think so, but I worry about this special case that was cpufamily 5 + lots of extensions (including cmov).
[12:52] <persia> That said, could be not so many folks have them, really.
[12:53] <persia> mvo, if nobody volunteers with testable hardware for my corner case, I'd go ahead with the rule you described: maybe we get another bug, maybe we don't.
[12:54] <mvo> yeah, it will certainly help a lot of people and that worst that can happen is that people need to stay whereas currently it will break for them
[12:54] <mvo> so we error on the safe side
[12:54]  * mvo adds this code
[12:55] <mvo> thanks persia!
[12:55] <mvo> and sladen
[12:55]  * persia points at sladen, having mostly gotten in the way with an untestable corner case
[13:03] <ogra_ac> stgraber, are you here ?
[13:08] <popey> persia: which specific eee pc, I thought they started with underclocked 900MHz celerons. Don't recall any having VIA CPUs
[13:08] <persia> It would have been the very first model: the second model switched from VIA to Intel.
[13:09] <persia> There is other C7M stuff out there (the Aigo MID, etc.), but I thought the Eee was the biggest seller.
[13:10] <persia> Mind you, within a couple months, Atom was released, and everyone switched to that, so I may be hoping to allow a nigh-insignificant number of folks to upgrade.
[13:10] <popey> I'm certain the first model (of which I had two) were 900MHz celerons underclocked to 600MHz. The model number was 701.
[13:11] <persia> OK.  I was also thinking of the 701s.  if you're sure they were celerons, then I misremember, and it doesn't matter, because all the other things I believe to have shipped with C7M required proprietary drivers for enough components that we've never supported them well.
[13:12] <popey> yes, other C7M devices were around at the same time
[13:12] <popey> I recall them being compared in performance tests
[13:15] <popey> can't imagine there's many of them about though.
[13:17] <persia> I hope there either aren't many, or I misremember the odd config so that they can't actually run maverick.
[13:17] <systemofadown> my pc has better performance than a 8-core pc in fabonacci
[13:18] <systemofadown> fibonacci?
[13:18] <cjwatson> the latter
[13:20] <ogra_ac> whats the prefix for preformance- specs ? a spec search for that term doesnt reveal anything
[13:21]  * ogra_ac wants to know if his spec fits the scheme 
[13:21] <cjwatson> I wouldn't have expected naïve implementations of Fibonacci's algorithm to parallelise; the traditional x[n] = x[n - 1] + x[n - 2] implementation has an inherent limit on parallelisability.  Obviously you can implement it using Binet's formula for the closed form and parallelise that way (should you care).
[13:22] <persia> ogra, "other"
[13:22] <ogra_ac> persia, really ? why is it listed as track ?
[13:23] <persia> Oh my.  It wasn't a track 6 hours ago.
[13:23] <ogra_ac> it was when i looked yesterday
[13:23] <ogra_ac> i just cant find any specs for it
[13:23] <persia> (or else I somehow didn't read it, and did notice the layout looked unbalanced, meaning there was a very strange blind spot)
[13:24] <cjwatson> goodness me.  two hours later, and the initial autosync queue flush is only up to luatex
[13:24] <persia> a-l in two hours is better than usual though.
[13:25] <cjwatson> this is just the flush step; I don't remember how long that usually takes.  the bit that normally takes ages is constructing the syncs
[13:25] <cjwatson> (because sync-source falls over whenever it encounters something that needs manual resolution, and then you have to start again from the beginning, so it has worst-case quadratic time
[13:25] <cjwatson> )
[13:25] <seb128> tkamppeter, hi, do you have a vcs for system-config-printer?
[13:25] <geser> cjwatson: is this with NEW packages or only existing ones?
[13:25] <cjwatson> existing
[13:26] <ogra_ac> pitti, did the alsa-utils SRU go through (i have some TI devs waiting for it)
[13:26] <ogra_ac> pitti, to proosed that is
[13:26] <ogra_ac> *proposed
[13:28] <seb128> tkamppeter, could you drop the depends on python-gnome2?
[13:29] <persia> ogra, https://launchpad.net/ubuntu/maverick/+queue?queue_state=3 is a handy way to see recent acceptances into -propsed
[13:29] <seb128> tkamppeter, https://code.edge.launchpad.net/~mterry/ubuntu/maverick/system-config-printer/gtkbuilder/+merge/37621 as well
[13:30] <persia> ogra, And according to https://launchpad.net/ubuntu/maverick/+queue?queue_state=1 it's still pending
[13:30] <ogra_ac> persia, no alsa-utils ...
[13:30] <ogra_ac> yeah
[13:31] <persia> Right.  Just want to give you tools to be able to say "hey pitti, could you approve that?" rather than "did you approve that?"
[13:34] <ogra_ac> persia, i did the former this morning, thats why i asked ;)
[13:34] <persia> Oh, so with the tools you could even have reached "Why didn't you approve it yet" :)
[13:34] <ogra_ac> heh
[13:59] <pitti> yeah, yeah, I'll get to it :)
[14:00] <ogra_ac> thanks :)
[14:00] <ogra_ac> i know you are busy spec writing ...
[14:03] <doko> cjwatson, pitti: any feedback/corrections on http://paste.ubuntu.com/513846/ ?
[14:05] <pitti> ogra_ac: actually not; I'm busy finishing my bug assignments on my current OEM project by UDS :)
[14:06] <pitti> doko: looks fine to me
[14:06] <ogra_ac> oh
[14:07] <ogra_ac> pitti, i'm just writing a spec for arm thin client implementations btw, might be helpful to have some input from you at UDS :)
[14:11] <werber> http://www.uk.spmgame.com/partner.php?ID=267046
[14:11] <werber> http://www.uk.spmgame.com/partner.php?ID=267046
[14:14] <pitti> is that the new guerilla marketing strategy?
[14:14] <ogra_ac> heh
[14:21] <cjwatson> doko: looks OK
[14:21] <cjwatson> doko: (I'll fix libdebian-installer for --no-as-needed, BTW)
[14:21] <cjwatson> and syslinux
[14:22] <cjwatson> and newt
[14:23] <pitti> ogra_ac: please reupload alsa-utils with -v; it covers two versions
[14:23] <pitti> (rejected)
[14:23] <ogra_ac> -v for the changelog you mean ?
[14:23] <ogra_ac> will do
[14:23] <pitti> right
[14:23] <doko> cjwatson: note, it's --no-add-needed
[14:28] <cjwatson> doko: sorry, yeah, I always get those confused
[14:31] <doko> cjwatson: me too :) it's now renamed to --no-copy-dt-needed-entries
[14:31] <doko> could somebody approve the email to u-d-a?
[14:32] <ogra> pitti, done
[14:33] <pitti> doko: looking
[14:35] <pitti> doko: flushed
[14:39] <benste> hi folks, could someone explain me which authentification is used for GDM Settings, software center and expanded user settings ? - I'm only getting a box asking to authentificate without the possibility to enter my pasword - but don't know where to start troubleshooting
[14:40] <mvo> persia: I just found a blog post by someone telling the world that his via c7m works with 10.10, so we should be fine there
[14:41] <persia> nenolod, You may find #ubuntu is a better place to ask.  This isn't a support channel.
[14:41] <persia> mvo, Well, actually, the opposite, because I think c7m is family5 + cmov, which I think fails the test.
[14:41] <persia> I might be wrong, but I remember having some issues with the ubuntu-mid images on c7ms when we did tests for i686.
[14:42] <persia> So I fear the SRU would prevent some folk from updating who otherwise might be supported.
[14:43] <mvo> persia: I got the cpuinfo from that cpu, it says family 6 so we are good. but who knows, there might be multiple versions of that cpu :/ I was not able to find a "repository" of cpuinfo output
[14:44] <persia> mvo, Then I'm completely wrong.  Sorry to have raised the issue.
[14:46] <mvo> persia: its all good (in fact, great) - I added a test-case for this to the u--m repository and to be honest, this kind of change makes me nervous, so any hint/blocker/issue should be followed and checked
[14:49] <persia> mvo, Heh, OK.  I feel better knowing that your increased peace of mind made up for your increased work :)
[14:49] <mvo> persia: peace-of-mind++
[15:04] <tkamppeter> seb128, I will have a look into it. What is the problem of python-gnome2?
[15:04] <seb128> tkamppeter, it's legacy old technology and system-config-printer stopped using it
[15:05] <seb128> tkamppeter, the depends is buggy, it just makes some old component still be pulled in for no reason
[15:08] <tkamppeter> seb128, OK. Is this only needed for Natty or also as SRU? The use of Glade is already removed upstream and Tim Waugh (upstream author) will also do some file renamings soon to not have wrong impressions. The removal of python-gnome2 and Glade dependencies is no problem for me.
[15:09] <seb128> tkamppeter, only natty
[15:09] <seb128> tkamppeter, thanks
[15:14] <cjwatson> doko: I can't reproduce the problem with libdebian-installer in https://wiki.ubuntu.com/RobSavoye/GoldFixes; it builds fine in an up-to-date natty chroot
[15:16] <doko> cjwatson: yes, maybe he didn't finish the check. found two others which didn't need fixing
[15:16] <cjwatson> doko: is the python3.1 sync on cocoplum yours?  if so you can flush it now, I'm done with the initial autosync (at least of main, would like to move on to contrib / non-free / other related stuff)
[15:17] <tkamppeter> seb128, so I remove python-glade2 and python-gnome2 from the dependencies of system-config-printer-gnome. Do I need to put any replacement or to remove more?
[15:18] <seb128> tkamppeter, grepping in the source you need to depends on
[15:19] <seb128> tkamppeter, on python-gnomekeyring
[15:19] <seb128> jobviewer.py:    import gnomekeyring
[15:19] <doko> cjwatson: done
[15:19] <cjwatson> ta
[15:20] <tkamppeter> seb128, thanks.
[15:20] <seb128> tkamppeter, you're welcome
[15:31] <tkamppeter> seb128, new s-c-p uploaded.
[15:31] <seb128> tkamppeter, thanks!
[15:32]  * pitti makes seb128 a honorary member of https://edge.launchpad.net/~ubuntu-cruft-busters
[15:33] <seb128> pitti, ;-)
[15:47] <chrisccoulson> lamont - you there?
[15:50] <cjwatson> doko: syslinux also works unmodified in natty, but at least newt fails
[15:51] <cjwatson> (missing -ldl by the looks of it)
[16:26] <cjwatson> doko: can I mark things as done directly in RobSavoye/GoldFixes?
[16:27] <doko> cjwatson: I think that would be good
[16:31] <cjwatson> pitti: can we disable the "no work items defined" mails until after UDS?
[16:31] <cjwatson> apparently some specs have been accepted prematurely
[16:32] <ogra_ac> just to be sure, i didnt recieve any release meeting mail or cancellation, there is none today, right ?
[16:33] <cjwatson> so I understand
[16:34] <ogra_ac> thanks
[16:34] <pitti> cjwatson: sure, no objection; I just have about zero time right now, need to fix an apt regression today which causes OEM builds to fail
[16:37] <cjwatson> pitti: what would be your preferred approach?  is a new configuration key overkill?
[16:37] <cjwatson> (happy to implement whatever)
[16:41] <pitti> cjwatson: we could just temporarily comment out the warning message about the "no WIs"
[16:44] <cjwatson> pitti: ok, fair enough - done
[16:44] <doko> ERROR: 90 tests were run,
[16:44] <doko> 71 failed unexpectedly.
[16:44] <doko> 8 tests were skipped.
[16:44] <doko> hmm, tar ...
[16:44] <cjwatson> er - really done now
[17:04] <cjwatson> initial autosync pass has finished for existing packages, FWIW
[17:05] <pitti> yay
[17:05] <pitti> cjwatson: not that many this time, I figure?
[17:06] <cjwatson> I didn't count, TBH
[17:06] <cjwatson> didn't seem too bad
[17:07] <sebner> cjwatson: 3 days waiting for a buildmachine is pretty normal so bad :P
[18:17] <SpamapS> ttx: I don't see a "web scale" blueprint yet, should I be adding that?
[18:26] <cjwatson> doko: are you going to merge debconf, or should I?
[18:26] <cjwatson> (happy to)
[18:27] <cjwatson> cody-somerville: you're marked as touched-it-last for base-installer due to that change for live-installer, but most of the changes there are mine - do you want me to handle the merge?
[18:28] <cody-somerville> cjwatson, You're more than welcome to do it if you'd like. I won't have time to get around to it today or early next week for sure.
[18:35] <cjwatson> cody-somerville: ok, no problem
[18:42] <ggeorgy> Hi !!! :D
[18:44] <ggeorgy> i installed ubuntu 10.10 and is great! But why the logout sound not work???
[18:44]  * hyperair points at #ubuntu
[18:45] <hyperair> actually it sounds like a potential bug. maybe you could report it on bugs.launchpad.net/ubuntu/+source/gnome-session (i think that's what's responsible for logout sounds)
[18:47]  * hyperair loves the way sabdfl phrased "odd super-self-interested muppet who expects me to singlehandedly make his wet dreams of technology kfuturism come true"
[18:48] <ggeorgy> the logout sound not worked also in  ubuntu 10.4 !!!
[18:49] <hyperair> ggeorgy: file a bug, please
[18:51] <ggeorgy> i think this bug alreay reported but the problem will not be resolve
[18:52] <hyperair> do you have the bug number?
[18:52] <ggeorgy> no
[19:17] <cjwatson> cody-somerville: ... you're lucky too, it's a conflict-resolution-and-a-half
[19:22] <ggeorgy> i reported that bug #661385
[19:24] <doko> cjwatson: please go ahead, I'm done for today
[19:25] <ggeorgy> ok
[19:29] <cjwatson> doko: ok, thanks
[19:30] <bilalakhtar> doko: hi
[19:30] <ttx> SpamapS: sure
[19:31] <kirkland> cjwatson: hi there;  i have a couple of questions about multiverse, if you have a second ...
[19:31] <kirkland> cjwatson: ahs3 has been working on building paravirtual drivers for Windows
[19:32] <kirkland> cjwatson: Windows guests running in Ubuntu KVM can install these PV drivers and take advantage of PV disks, network, and memory ballooning
[19:32] <kirkland> cjwatson: the source code for the drivers themselves are GPL
[19:32] <kirkland> cjwatson: but we're not able to cross compile them on Linux
[19:32] <kirkland> cjwatson: ahs3 has to build them in a Windows VM
[19:33] <kirkland> cjwatson: the output is a little ISO (11MB, I think)
[19:33] <kirkland> cjwatson: we're trying to decide the best way to make this ISO available
[19:33]  * ahs3 listens attentively...
[19:34] <kirkland> cjwatson: one way would be through an Ubuntu package, that drops the ISO somewhere reasonable in /usr/share, and then the user can make that available to guests
[19:34] <kirkland> cjwatson: by hot-add of a virtual cdrom to the guest
[19:34] <kirkland> cjwatson: alternatively, we could, perhaps, just publish the ISO somewhere on ubuntu.com, or in the wiki or documentation
[19:35] <kirkland> cjwatson: I was looking to bounce this off of you (or someone) first, though
[19:35] <cjwatson> I'm trying to understand what your question about multiverse is
[19:38] <cjwatson> oh, you're asking whether you can put binaries built ex-archive from GPLed software in multiverse?
[19:38] <SpamapS> sounds to me like it would have to have  Build-Depends: windows
[19:38] <kirkland> cjwatson: yes, right
[19:38] <ogra_ac> sounds ok imho
[19:38] <kirkland> cjwatson: so we can publish the source, but the buildd's ain't gonna build it
[19:38] <ScottK> We have plenty of binary only stuff in Multiverse.
[19:38] <cjwatson> ogra_ac: he asked me
[19:39] <ogra_ac> cjohnston, yes
[19:39] <ogra_ac> err cjwatson
[19:39] <cjwatson> I think you should ask a lawyere
[19:39] <cjwatson> *lawyer
[19:39] <kirkland> cjwatson: alrighty
[19:39] <kirkland> cjwatson: i'm really not that concerned about the legal distribution question -- i think we're clear
[19:40] <kirkland> cjwatson: i'm more wondering about the ubuntu policy on such matters
[19:40] <cjwatson> binary-only in multiverse is fine, but whether you can distribute GPLed binaries when I'm not sure we can distribute "the scripts used to control compilation and installation of the executable" (GPLv2 s. 3) ...
[19:40] <kirkland> cjwatson: something that's built once somewhere
[19:40] <kirkland> cjwatson: gotcha;  ahs3: do you have any more info about this part?
[19:41] <kirkland> ahs3: ie, the licensing of the source and the required build tools
[19:41] <ahs3> kirkland: the source is _all_ GPL2
[19:41] <cjwatson> how far down does "scripts" have to go, and are the resulting executables contaminated by linking with system components?
[19:42] <cjwatson> as in is there something like libgcc that gets statically linked into the resulting binaries?
[19:42] <ahs3> there are components of the Windows DDK (Driver Development Kit) that are included, and are covered by a freely redistributable clause in the DDK
[19:42] <cjwatson> freely redistributable doesn't meet the terms of the GPL
[19:43] <cjwatson> hence - ask a lawyer, this isn't a field I'm used to
[19:43] <cjwatson> more used to dealing with proprietary code on top of a free OS
[19:43] <kirkland> cjwatson: okay, thanks
[19:43] <ahs3> correct.  there's a note in the DDK about an exception to the GPL that's been approved
[19:43] <kirkland> ahs3: we'll need to talk to amanda brock, canonical legal counsel
[19:43] <SpamapS> Yeah, I'd bet this needs to be LGPL'd if it links any software like that.
[19:44] <ahs3> i have to go look it up again, but the RH folks had this covered in the docs in the source
[19:45] <cjwatson> so do you mean that the source code is actually GPL plus special exception for the Windows DDK?  An exception in the Windows DDK licensing wouldn't be enough
[19:45] <cjwatson> if that's the case, it would be OK in multiverse; however, distributability aside, I wonder if a package is really the best way to distribute this
[19:45] <ahs3> afaiui, the former.  agreed, the DDK licensing would not be sufficient
[19:45] <cjwatson> as in most convenient
[19:46]  * SpamapS likes the idea of an iso that is hot-added to the VM as a CD-ROM
[19:46] <cjwatson> if you're integrating it tightly in other packages, then a package would be reasonable
[19:46] <SpamapS> thats definitely how the VMware folks prefer to distribute guest enhancements.
[19:47] <SpamapS> and it works really nicely
[19:47] <ahs3> hrm.  these would only be useful pre-installed in a windows guest -- which we obviously can't redistribute
[19:48] <SpamapS> ahs3: they can't be installed after the fact?
[19:48] <ahs3> that's the only tight integration that makes sense
[19:48] <SpamapS> like, during install you get the slow SCSI emulation.. but after the drivers are in, you get the hotness?
[19:48] <ahs3> SpamapS: they can, but not automagically.  there are manual steps involved :(
[19:49] <ahs3> right.  you get the slow stuff during install.  shut down your VM, restart it with the virtio incantations, then update the drivers in the guest
[19:51] <ahs3> all the ISO does is allow you to attach it to the VM easily so you can get the "new" drivers
[19:51] <SpamapS> Doesn't windows have some kind of "load disk drivers from a CD" mechanism that could be leveraged?
[19:52] <ahs3> possibly.  that's way past my windows expertise, tho....
[19:54] <achiang> ahs3: SpamapS: i've definitely seen that option in the windows installer
[19:55] <SpamapS> I know win2k had it as I had a very new server with a totally unsupported RAID and was able to install using the extra CD-ROM from the mfgr.
[19:55] <SpamapS> so thats even more compelling an argument for the ISO
[19:55] <SpamapS> and if possible, formatting it as a windows driver disk
[19:56] <SpamapS> which, it may already be, as the same thing would be necessary for adding the driver manually later
[19:56] <Isaick> Is there an easy way to add/remove environment variables when creating an install package?
[19:57] <ahs3> SpamapS: yeah, i don't think there's any special format for a driver disk.
[19:57] <cjwatson> Isaick: our policy manual says that programs distributed in packages must not depend on particular environment variables being set in order to work
[19:57] <cjwatson> Isaick: why do you need this?
[19:58] <ahs3> SpamapS: so, if we just dropped the iso in /usr/share/blahblah, you'd find it useful?
[19:58] <Isaick> cjwatson: I've created a python app that I plan on installing in /opt/.  I need to be able to append the /opt/[app] directory to the PYTHONPATH variable.
[19:58] <cjwatson> Isaick: you should either distribute a wrapper script that sets the environment variable, or you should make your top-level Python program prepend to sys.path
[19:58] <cjwatson> setting PYTHONPATH for everything isn't the right fix for that
[20:01] <cjwatson> Isaick: if you lay out your files right, you may not even need that - for example, one package I maintain has its modules in /usr/lib/germinate which isn't on sys.path.  however, I don't need to set PYTHONPATH or modify sys.path to make this work; I just make /usr/bin/germinate be a symlink to /usr/lib/germinate/germinate.py
[20:02] <cjwatson> this may or may not be appropriate
[20:02] <Isaick> cjwatson: I'll take a look.  Thanks for the info
[20:08] <kirkland> jdstrand: ping
[20:09] <jdstrand> kirkland: yes?
[20:10] <kirkland> jdstrand: wondering if you had a couple of minutes to de-NEW a package to universe for me
[20:10] <jdstrand> kirkland: what package?
[20:11] <kirkland> jdstrand: https://edge.launchpad.net/ubuntu/natty/+queue?queue_state=0&queue_text=bikeshed
[20:13] <jdstrand> kirkland: sure
[20:13] <kirkland> jdstrand: awesome, thanks
[20:14] <psychicist> hi, I know this isn't a support channel but I have a question pertaining to ubuntu 10.4(.1) i386
[20:14] <psychicist> today I have had two issues with people calling me and saying that they lost sound
[20:15] <psychicist> in one case it was solved by applying updates and in the other case sound has stopped working after applying updates
[20:15]  * ScottK notes you didn't actually ask a question.
[20:15] <psychicist> so I was wondering if anyone in the ubuntu teams is aware of these problems and what to do about them
[20:16] <ScottK> Generally we are very interested to find out about post-release updates causing regressions.
[20:17] <ScottK> Do you have access to details of the system that lost sound/know what package updates they installed?
[20:18] <psychicist> I am typing on that system now, is there a place where I could find that information?
[20:18] <psychicist> I see /var/cache/apt/history.log, is that what you need?
[20:19]  * hunger just finished the first round of natty upgrades:-)
[20:19] <ScottK> I'm not the best person to answer sound questions.
[20:21] <psychicist> okay, is there anyone who does know more about the sound system or a mailing list where I could contact the person(s) responsible for the sound system?
[20:21] <ScottK> psychicist: Open a shell and type "ubuntu-bug".
[20:21] <ScottK> It will give you point and click options to specifically collect information for a sound related bug.
[20:22] <ScottK> Once you're done.  Give us the bug number here.
[20:33] <SpamapS> hmm, so natty is open.. when does 'do-release-upgrade -d' start working?
[20:33] <kklimonda> around alpha?
[20:34] <SpamapS> ah
[20:34]  * SpamapS dist-upgrades ;)
[20:35] <Quintasan> barry: ping
[20:35] <kirkland> Spads: sudo sed -i s/maverick/natty/g /etc/apt/sources.list ;-)
[20:35] <kirkland> SpamapS: ^
[20:35] <kirkland> Spads: sorry
[20:36] <psychicist> ScottK: it took me some time to fill in all the information, but here it is #661427
[20:36] <ScottK> psychicist: Thanks.
[20:36] <SpamapS> kirkland: :%s/maverick/natty/g in vim about the same time as you sent that. ;)
[20:36] <ScottK> Bug 661427
[20:36] <kirkland> SpamapS: heh
[20:37]  * SpamapS will wait for mongodb to finish building before he goes any further though.
[20:37] <barry> Quintasan: pong
[20:38] <ScottK> !regression
[20:38] <SpamapS> doh.. filled my vm's disk..
[20:38] <ScottK> Maybe that's not the right one.
[20:38] <Quintasan> barry: I'm trying to update SIP package and I want to send it to Debian, I was wondering in which section in rules I should put dh_python3
[20:39] <ScottK> !regression-alert
[20:39] <ScottK> See Bug 661427 for details.
[20:39] <ScottK> psychicist: ^^^ That should get someone who knows better than I do to look at it.
[20:40] <slangasek> ScottK, psychicist: what update caused the regression?
[20:40]  * ScottK looks at psychicist.
[20:40] <slangasek> this is filed against alsa-driver, but I don't see an alsa update in the attached log
[20:40] <ScottK> (all I know is he showed up with a regression)
[20:40] <slangasek> or even anything sound-related
[20:41] <ScottK> slangasek: That's the default when you file a bug about sound.
[20:41] <slangasek> sure
 today I have had two issues with people calling me and saying that they lost sound
 in one case it was solved by applying updates and in the other case sound has stopped working after applying updates
[20:41] <ScottK> ^^^ is what he said when he arrived.
[20:42] <barry> Quintasan: i'm actually not sure, i haven't done one of those yet.  if ScottK doesn't know off the top of his head, i will find out
[20:42] <psychicist> slangasek: I have attached a log of the updates I applied on this system just a few hours ago, I don't know if it was alsa-driver or something else, because as far as I know nothing sound-related was updated
[20:42] <psychicist> but sound has stopped working nonetheless
[20:43] <slangasek> psychicist: when did sound stop?  Immediately after applying the updates?  After a reboot?
[20:44] <Quintasan> barry: well, the debian's wiki says it should be before any dh_strip but I can't find any direct dh_strip, well I'll just put it before them and see what happens
[20:44] <psychicist> slangasek: I am not sure about that, I rebooted after applying these updates
[20:45] <slangasek> psychicist: I see no indication that this regression is related to the latest round of updates.  When was your last reboot prior to this?  Also, please attach /var/log/apt/term.log to the bug report (I've followed up to the bug with a request for this information)
[20:45] <slangasek> afk, back shortly
[20:45] <barry> Quintasan: let me know how it goes :)
[20:45] <ScottK> Quintasan: If there's a direct reference to dh_pysupport, put it just after that.
[20:47] <Quintasan> ScottK: okay
[20:47]  * Quintasan hopes the universe won't explode after pbuild
[20:53] <jdstrand> a/win 14
[20:53] <barry> Quintasan: here's an example from markupsafe which has been converted to dh_python2 and dh_python3: http://pastebin.ubuntu.com/514095/
[20:54] <ScottK> (however we don't want to pysupport -> dh_python2 in Ubuntu without concurrence from the Debian maintainer)
[20:55] <Fitzsimmons> how do you add a menuitem to an existing appindicator? I'm pretty sure gwibber and pidgin are doing this, but I can't find the code for it
[20:55] <barry> ScottK: right.  hopefully soon :)
[20:56] <Quintasan> barry: well, it seems it's read to go, can you paste me the control file somewhere as well?
[21:03] <kirkland> jdstrand: thanks
[21:12] <barry> Quintasan: http://pastebin.ubuntu.com/514103/
[21:12] <psychicist> ScottK: slangasek: thanks for your help, I will try a reinstall on another partition on Sunday and see if the problem can be reproduced
[21:40] <sistpoty> resizing an ext4 fs: 5$, resizing lvm underneath it: 10$, recovering data from backup: priceless :)
[21:44] <SpamapS> sistpoty: :)
[21:44] <nemo> I know this has come up before.
[21:44] <nemo> (and there are a few bugs on it)
[21:44] <nemo> buuuut
[21:44] <nemo> # ls -lh .xsession-errors
[21:44] <nemo> -rw------- 1 1000 1000 373G 2010-08-26 13:11 .xsession-errors
[21:44] <Quintasan> barry: thanks, my internet is slower than a turtle crawling today
[21:44] <nemo> just throwing it out there :)
[21:45] <Quintasan> >373G
[21:45] <sebner> huhu sistpoty =D
[21:45] <Quintasan> No, how did you manage to get it so big? :O
[21:45] <nemo> Quintasan: good question
[21:45] <nemo> I'm trying to figure that out
[21:45] <nemo> unfortunately, seeking through a 373GiB file takes a while
[21:45] <nemo> dd seek=381931 bs=1M count=1 .xsession-errors
[21:45] <Quintasan> Why not just rm -rf it?
[21:46] <nemo> Quintasan: I want to see what is in it before I erase it
[21:46] <nemo> Quintasan: this was on an ubuntu machine that was left idle for a while
[21:46] <Quintasan> >a while
[21:46] <nemo> it disturbs me that this sort of thing could happen to a machine just sitting there
[21:46] <nemo> Quintasan: a while like.... a few weeks to a month
[21:46] <nemo> or so
[21:46] <Quintasan> ...
[21:46] <Quintasan> @_@
[21:46] <Quintasan> 383GB of text in a month?
[21:47] <Quintasan> I don't think I can even download that much
[21:47] <nemo> regardless, xsession-errors should have some sort of cap
[21:47] <nemo> like any log file
[21:47] <Quintasan> -rw------- 1 quintasan quintasan 1,5M 2010-10-15 22:47 .xsession-errors
[21:47] <nemo> rolling limit of, oh, 5 megs or something
[21:47] <nemo> not sure if ubuntu has any control over that
[21:47] <nemo> Quintasan: I know. that's not atypical
[21:47] <nemo> mine at home is less than a meg
[21:48] <nemo> with an uptime of 26 days
[21:48] <nemo> but the fact that it happens at all...
[21:48] <barry> Quintasan: no worries
[21:49] <nemo> man. I wish dd could tell me how far it had seeked into the file.
[21:49] <kklimonda> sending USR1 signal doesn't show that?
[21:49] <nemo> nope
[21:49] <nemo> not for seek
[21:50] <nemo> does tell me it has been running for 52 minutes now
[21:50] <kklimonda> :)
[21:50] <sistpoty> hi sebner
[21:51] <nemo> hmmmm
[21:51] <nemo> you know
[21:51] <nemo> maybe I should have just started from the start
[21:52] <nemo> Starting Oracle Universal Installer... Checking installer requirements... Checking Temp space: must be greater than 400 MB. Checking monitor: must be configured to display at least 256 colors    Failed Could not execute /usr/X11R6/bin/xdpyinfo Checking if CPU speed is above 300 MHz.    Actual 2992 MHz    Passed
[21:52] <nemo> etc etc
[21:52] <nemo> Some optional pre-requisite checks have failed (see above). Continue? (y/n) [n] Continue? (y/n) [n] Continue? (y/n) [n] Continue? (y/n) [n] Continue? (y/n) [n]
[21:52] <Quintasan> Oh god...
[21:52] <nemo> repeat Continue? (y/n) [n]
[21:52] <nemo> for line after line after line
[21:53] <Quintasan> tail -n 50>
[21:53] <nemo> TBH I'm not sure why the "Oracle Universal Installer" is running - much less as this user
[21:53] <Quintasan> ?
[21:53] <nemo> Quintasan: tried that first, that hung forever so I decided to try dd w/ a large block size
[21:53] <nemo> aaaanyway, a process spamming STDERR should not fill up a disc like this :-p
[21:54] <Quintasan> No wonder, no normal text processing expects 300+ GB file
[21:54] <Quintasan> text processing app*
[21:58] <Quintasan> nemo: a request from #ubuntu-pl -> "Pastebin it ;)"
[21:58] <nemo> haha
[21:59] <nemo> here we have a clue
[21:59] <Quintasan> Years worth of reading
[21:59] <nemo> the .xsession-errors has this not far from the continues.
[21:59] <nemo> near the top of the file
[21:59] <nemo> Wed Aug 25 15:54:02 2010
[21:59] <nemo> the last date in the file?
[21:59] <nemo> -rw-------  1 1000 1000 400485761024 2010-08-26 13:11 .xsession-errors
[22:00] <nemo> so. in about 3 days. it generated 373GiB of spam
[22:00] <nemo> if it really is those 20 characters, repeated...
[22:01] <nemo> that's around 75k repeats of the message in the log
[22:01] <nemo> per second
[22:01]  * nemo rechecks his math
[22:02] <Quintasan> no, seriously, this is some sort of a record
[22:02] <nemo> yep. 75k or so repeats per second
[22:02] <nemo> I'm astounded the machine functioned
[22:03] <nemo> I'm guessing some incredibly stupid shell script in that installer, running in a loop,
[22:03] <nemo> I wonder if by any chance that was actually part of an attempted install
[22:03] <Quintasan> at least you know that your machine is powerful when it comes to printing the text to a file :P
[22:04] <nemo> you know. there are a few bugs on this issue...
[22:05] <nemo> but I'm thinking this one might give it a little more profile
[22:05] <nemo> there really is no reason whatsoever this file should have gone beyond even a few megabytes before starting to truncate
[22:05] <Quintasan> nemo: post it on planet
[22:05] <nemo> planet?
[22:05] <Quintasan> Ubuntu Planet
[22:05] <Quintasan> You are not a member yet?
[22:05] <nemo> maybe. dunno
[22:05] <nemo> I don't use that stuff very often
[22:06] <nemo> woah. wait
[22:06] <nemo> I fail at math
[22:06] <nemo> or rather, calnedars
[22:07] <jpds> Doesn't sound like it's Ubuntu's fault.
[22:07] <nemo> that was from the 25th at 16h to the 26th 13h - really it was 21¼h or so
[22:07] <nemo> jpds: well. I suppose in the sense that .xsession-errors behaviour is not ubuntu's decision
[22:08] <nemo> ok. that was really 256KiB/s written
[22:09] <Quintasan> nemo: Why don't you release it as a book? :D
[22:09] <nemo> it'll compress really well
[22:10] <nemo> jpds: doesn't mean ubuntu couldn't choose to add truncation
[22:10]  * nemo looks into what options there are for that file
[22:13] <nemo> MAX_XSESSION_ERROR_BYTES
[22:13] <nemo> hmmm
[22:14] <nemo> maybe it *is* ubuntu's fault :)
[22:14] <nemo> http://ubuntuforums.org/showthread.php?t=980958 - clearly this wasn't the only case
[22:14] <nemo> ah. gdm does its own purging
[22:16] <nemo> looks like this problem spans distros. no one has offered a decent solution
[22:16] <Quintasan> 400 505 700 352 <- I counted the number of characters in this file, assuming each is char is 1 bit
[22:17] <nemo> er. you mean 1 byte right
[22:17] <Quintasan> oh yes
[22:17]  * Quintasan can never get this right
[22:17] <nemo> also, the number of characters is presumably same as the number of bytes. divide by 20 and you get how many times that continue message shows up in there
[22:18] <nemo> thus my estimate of 256,000 repeats of the message each second while that log was active
[22:18] <nemo> looks like bugs have been filed on this for over 6 years, with no solutions offered
[22:18] <nemo> so clearly this one is not going to be solved tonight
[22:18] <Quintasan> :O
[22:18] <Quintasan> nemo: can you compress it?
[22:19] <nemo> heh
[22:19] <nemo> probably
[22:19] <Quintasan> or just tell me how long it would take to compress it :D
[22:20] <nemo> jpds: IMO the simple solution would be for xinit to change from:
[22:20] <nemo> looks like bugs have been filed on this for over 6 years, with no solutions offered
[22:20] <nemo> oops
[22:20] <nemo> er. to continue
[22:20] <nemo> to change from:
[22:20] <nemo> exec >"$ERRFILE" 2>&1
[22:20] <nemo> to:
[22:20] <nemo> exec 2>&1 | head -c 5000000 > "$ERRFILE"
[22:21] <nemo> then after the 500,000
[22:21] <nemo> presumably the rest would just disappear into the void
[22:21] <nemo> er. the 5,000,000
[22:21] <Quintasan> >probably
[22:21] <nemo> I believe I will file a bug asking if this would be feasible
[22:21] <Quintasan> attach your log as proof :P
[22:22]  * Quintasan is curious how long it would take to compress this
[22:27] <nemo> https://bugs.launchpad.net/ubuntu/+source/x11-apps/+bug/661494
[22:28] <nemo> hopefully that's an appropriate place to file it
[22:28] <Quintasan> well, it's time to go to bed
[22:28] <Quintasan> night
[22:29] <Quintasan> nemo: well, I hope you get some response on this bug :)
[22:29] <nemo> me too :)
[22:30] <plasmasolutions> Hi@allI've got a problem: I used the gconf-sharp in a project some time ago...and I'd like to continue development but realized that this assembly is not packaged with ubuntu any more.. What's the appropriate replacement for this?
[22:35] <plasmasolutions> Maybe it's assembled in another assembly?
[22:35] <nemo> well. dd has been running for 1h 40m.  I think I'm going to give up on reaching the end of this file
[22:35] <nemo> I'm going to assume it consists of exactly what I suspect it does
[22:36] <nemo> a cat of the file just scrolled that endlessly until I got bored
[22:36] <ajmitch> plasmasolutions: it's still there, hasn't been removed at all
[22:38] <ari-tczew> please any core-dev open task on lucid in bug 626379
[22:42] <directhex> um... does anyone else experience CRAZY cpu load from dbus-daemon in maverick?
[22:45] <nemo> directhex: not personally, but have you tried ltrace -S or similar to see what it is doing?
[22:46] <plasmasolutions> Another question: Does anyone knows if it's possible to change another users gconf entry through gconf-sharp?
[22:46] <plasmasolutions> Specifically in this case: changing the background image of gdm through gconf-sharp
[22:47] <Pilif12p> Hey guys. Are there any bindwood developers in here? If so, doesb http://crash-stats.mozilla.com/report/index/bp-9f94cd73-56c7-4381-b09a-86a2f2101014 look like it's related to it?
[22:49] <directhex> nemo: i'll try that next time
[22:49] <directhex> problem is, when dbus misbehaves, yer screwed. try restarting it without doom
[22:54] <nemo> directhex: I've restarted it before
[22:54] <directhex> kills gdm, for one
[22:54] <nemo> that's doom? :)
[22:54] <plasmasolutions> Anyone a hint? Or idea how to do it?
[22:55] <nemo> what irritates me about dbus, personally, is it moderately complicates cron jobs
[22:55] <nemo> plasmasolutions: personally I have a loathing for the language, so really can't be of much help :)
[22:55] <nemo> (no familiarity)
[22:58] <directhex> plasmasolutions: this is a channel for development of, not on, ubuntu. try #mono on gimpnet
[22:58] <sistpoty> ari-tczew: is it a problem under lucid in the first place?
[22:59] <ari-tczew> sistpoty: I don't understand the question.
[23:00] <sistpoty> ari-tczew: you want a lucid task, right? my question is, if the bug is found in lucid as well (still reading the bug report though)
[23:00] <ari-tczew> sistpoty: yes, it exist in lucid.
[23:02] <sistpoty> ari-tczew: ok, thanks, task opened
[23:03] <ari-tczew> sistpoty: thanks
[23:43] <cjwatson> Pilif12p|afk: it might be worth trying #ubuntuone, since those guys work on bindwood