[01:14] <FireRabb1t> hey... i just accidently uploaded a packge to upload.ubuntu.come that REALLY should not have been uploaded
[01:14] <FireRabb1t> where do these files go / how can I get this deleted?
[01:22] <FireRabb1t> alright according to https://wiki.ubuntu.com/UbuntuDevelopment, anything with an unknown key is just silently ignored.. is there anyone here from Canonical that can confirm this?
[01:23] <azeem> that's the case, yes
[01:23] <FireRabb1t> do the files get deleted?
[01:23] <azeem> I don't know about that
[01:23] <azeem> I assume at some point they will
[01:24] <FireRabb1t> azeem: do you know who in here would know?
[01:25] <azeem> if they know, they will answer
[01:25] <FireRabb1t> fair enough, thanks :)
[02:21] <asac> FireRabb1t: should be no problem if your key hasn't the power to upload
[02:22] <asac> FireRabb1t: of course only if its an accident ;)
[02:23] <FireRabb1t> yea, it was. forgot the 'local' on 'dput'. :)
[04:00] <mekius> hey, working on a custom live cd here and have a few questions
[04:01] <mekius> does the live cd get copied to the hard drive upon install or does the installer use dpkg to install packages?
[04:01] <LaserJock> the squashfs gets copied
[04:02] <LaserJock> there aren't any .debs on the live cd
[04:02] <mekius> ok, i thought so
[04:02] <mekius> so whatever is on the live cd is pretty much what will be on the installed system?
[04:02] <mekius> obviously with a different init
[04:03] <LaserJock> some packages are removed after it's copied
[04:03] <mekius> ah ok
[04:04] <mekius> are there any plans to make customization any easier, or any advice on being able to see changes without having to rebuild an iso and load it up in qemu each time?
[04:04] <LaserJock> mekius: well, there is Reconstructor and Ubuntu Customization Kit GUIs for it
[04:05] <LaserJock> mekius: and for sure there are some people that would like to make it easier
[04:05] <LaserJock> but I don't think there is anything specifically in the works right now
[04:05] <mekius> yeah, but they require rebuilding the iso everytime and testing with qemu, then doing it all over again heh
[04:05] <LaserJock> well
[04:05] <mekius> ok
[04:05] <LaserJock> you don't know exactly what you're going to get until you built the .iso
[04:05] <mekius> i did see a wiki entry about some people with suggestions, but don't know if anyone started anything
[04:06] <LaserJock> we even do that for Ubuntu disks
[04:06] <mekius> well i know with archie, an arch based live cd, they offer a tool that will take a partition and make a live cd out of it
[04:06] <LaserJock> we build the .isos and test, and if there's a bug or something we rebuild new ones
[04:06] <mekius> so you can just customize it while it runs and then just pack up an iso
[04:06] <LaserJock> well, you can kinda do that
[04:07] <LaserJock> you mount the .iso loopback
[04:07] <LaserJock> and chroot in
[04:07] <LaserJock> and you can make changes
[04:07] <LaserJock> and then get out and repack the squashfs and .iso
[04:07] <mekius> yeah, right now i've got it unpacked
[04:08] <mekius> and i chroot in and do some karate and kick shit around, then exit and build a new iso
[04:08] <mekius> run it in qemu to test it, then rinse and repeat
[04:08] <LaserJock> yep, sound about right
[04:09] <mekius> alright, just sucks cause i can't see the effect of alot of my changes until i boot the iso and see something is screwed heh
[04:09] <mekius> i know it's not easy :)
[04:09] <mekius> sorry for the rant here heh
[04:09] <LaserJock> yep, that's pretty much how the Ubuntu devs do it too
[04:10] <mekius> jumped on to a project and found this method very time consuming :)
[04:10] <LaserJock> that's about the most easy going rant we've had in here fore a while ;-)
[04:10] <mekius> :)
[04:10] <LaserJock> I do know there are other people interested in that kind of thing
[04:10] <mekius> i try not to completely bash anyone for what they do, i'm sure this works fine for doing small updates, especially when you know the system well
[04:10] <LaserJock> it's just getting time and resources together to get something put together
[04:10] <mekius> just hard for someone starting out trying to do things quickly :)
[04:11] <mekius> yep, i hear ya
[04:11] <mekius> resources are always an issue in open source :)
[04:12] <mekius> how many people here work for Canonical?
[04:13] <LaserJock> I think there's gotta be like 20 or so people for Ubuntu development
[04:13] <mekius> k
[04:13] <LaserJock> total
[04:13] <mekius> thats a pretty good amount :)
[04:13] <LaserJock> well, depends
[04:14] <LaserJock> I think it's much smaller than say Red Hat or Novell
[04:14] <LaserJock> but having paid people definitely helps
[04:14] <mekius> oh sure
[04:14] <LaserJock> we have around 80-100 total devs
[04:15] <mekius> red hat and novell are alot more dug in
[04:15] <mekius> they have a lot of contracts to take care of
[04:16] <mekius> LaserJock: you do alot of dev?
[04:16] <LaserJock> mekius: I do what I can, which unfortunately isn't as much as I'd like :-)
[04:17] <mekius> yeah :)
[04:17] <LaserJock> I work mostly on Edubuntu these days
[04:17] <mekius> ah ok
[04:17] <mekius> good project, get em while their young :)
[04:18] <LaserJock> heh, yeah
[04:18] <mneptok> LaserJock: it's >20 on the distro team
[04:18] <LaserJock> mneptok: >30?
[04:18] <LaserJock> I remember when there was 15, but they keep adding people
[04:19] <mneptok> LaserJock: dunno. i can't keep track of new employees, either.
[04:19] <mneptok> but i know it's >20
[04:19] <LaserJock> LP gets the lions-share
[04:20] <LaserJock> but still, Debian's got ~1000 devs
[04:20] <LaserJock> gentoo's got 300+
[04:20] <LaserJock> I'd love to see numbers for RH and Novel
[04:20] <kylem> redhat has more people working on their kernel than canonical has engineers
[04:20] <kylem> (exaggeration, but still)
[04:21] <mneptok> kylem: but none of them have your scintillating criminal record.
[04:22] <LaserJock> there must be something about "bad boy" kernel hackers ;-)
[04:22] <kylem> mneptok, shh, i had that pardoned.
[04:25] <mneptok> pardoned != expunged
[04:25] <mneptok> you was robbed
[04:50] <whiprush_> LaserJock: ping
[04:51] <LaserJock> whiprush_: pong!
[04:51] <whiprush_> LaserJock: dude, you on im?
[04:51] <LaserJock> whiprush_: how are you?!
[04:51] <whiprush_> let's chat
[04:52] <LaserJock> IRC or jabber?
[04:52] <whiprush_> jabber
[04:52] <whiprush_> jorge.castro@gmail.com
[06:10] <mike>  I am a Vista 64bit Ultimate user, and I would like to transfer over to linux, but so far in 7.04 64bit and 32 dont mesh.  What do you guys recommend?
[06:11] <desrt> mike; you're in the wrong channel for this type of question
[06:12] <mike> Which channel then?
[06:12] <desrt> mike; but if i had to make a recommendation it would be to just run 32bit.  there are not enough compelling reasons to use 64bit for normal situations
[06:12] <mike> I just figure I ask developers straight up.
[06:12] <desrt> #ubuntu
[06:12] <desrt> see topic ^^.  no support question :)
[06:12] <mike> desrt, who is 64 bit for at this time?
[06:13] <desrt> mike; people who don't mind the disadvantages of running 64bit
[06:13] <desrt> basically, anything that comes in binary form (flash player, some video codecs, etc) won't work at all
[06:13] <desrt> since there are only 32bit versions of these binaries around
[06:13] <mjg59> flash can be made to work with minimal difficulty
[06:14] <mjg59> And most of the interesting video codecs are supported by ffmpeg
[06:14] <desrt> out of process linked against 32bit libc?
[06:14] <mjg59> desrt: Yeah
[06:14] <desrt> interesting
[06:14] <mjg59> nspluginwrapper
[06:14] <mike> except mpeg avc encoding!
[06:14] <mjg59> Has the handy side effect that it's out of process
[06:14] <desrt> mjg59; are there any good reasons to actually use 64bit, though?
[06:14] <desrt> mjg59; i find flash doesn't usually cause me trouble.  java is another story.
[06:15] <mjg59> desrt: You can use it to out of process stuff on the same architecture, too
[06:15] <mjg59> mike: If you want to use stuff where there's no 64-bit codec, then right now 32-bit is a much easier choice
[06:15] <ScottK> desrt: If you are doing CPU intensive processing using software optimized for 64 bit.
[06:15] <desrt> mjg59; i'd have imagined that to be the case :p
[06:15] <mjg59> desrt: Mildly better performance, access to more than 3GB of RAM
[06:16] <desrt> i'd imagine performance to be mildly worse under common workloads
[06:16] <mjg59> Nothing especially compelling. I'm doing it so I'm encouraged to fix the bugs
[06:16] <desrt> the whole 8-byte-pointer thing is a pretty annoying hit to take
[06:16] <mjg59> Basically every benchmark ever has shown x86_64 to be slightly faster than ia32
[06:16] <mike> I know this is a stupid question.  But in my Vista Ultimate I can still run 32bit apps on the fly with no hiccups.    I know you probably can't easily answer this.   But why cant my 32bit linux apps play nice inside a 64bit nix OS?
[06:16] <desrt> are they benchmarking integer math performance or something?
[06:17] <desrt> mike; they absolutely can
[06:17] <mjg59> mike: They can, but we don't have the infrastructure to make this easy
[06:17] <desrt> mike; you can run 32bit linux apps out of the box on ubuntu amd64
[06:17] <mjg59> desrt: The extra registers help
[06:17] <desrt> mike; but you need to have 32bit libraries to go with them.... and those don't currently ship as part of the 64bit version of ubuntu
[06:17] <desrt> mjg59; ahh.  good call.
[06:17] <mike> I got Intel CD2 at 3.2ghz with 64 bit Ubunut currenlty instaleld.
[06:17] <desrt> mjg59; i imagine the ip-relative addressing speeds up shared libraries too
[06:18] <desrt> no more bx-setting thunk
[06:18] <mjg59> mike: The difficulty in running 32-bit applications on 64-bit ubuntu is a failure on our part
[06:19] <mjg59> But I don't think anyone's actively working on it right now
[06:19] <desrt> the best solution is going to involve changes to apt/dpkg
[06:19] <mjg59> It's all been architected, just not implemented
[06:19] <mjg59> Tollef wrote his thesis on it, IIRC
[06:19] <desrt> i guess there would need to be two different sorts of dependencies
[06:20] <desrt> same-arch (for libraries) and any-arch (for stuff that is merely shelled out to)
[06:20] <mjg59> We already have that, effectively
[06:20] <desrt> do tell?
[06:20] <mjg59> Oh, I see what you mean
[06:20] <mjg59> Yeah, it would only special case libraries
[06:21] <desrt> on the other hand you could probably do it in the build infrastructure
[06:21] <desrt> and generate 3264 versions of all libraries
[06:22] <desrt> for arch "amd64" but containing 32bit code installed to /lib32
[06:23] <mike> So apt/dpkg    is only set up currently for one or the other?
[06:23] <desrt> with some very rare exceptions, yes
[06:23] <desrt> you either have a x86 system or an amd64 system
[06:23] <Amaranth> looking for a kubuntu person :)
[06:23] <desrt> amd64 has some very rare "32bit version" packages like libc
[06:24] <Amaranth> need to get the patch from http://bugs.kde.org/show_bug.cgi?id=147557 in
[06:24] <ubotu> KDE bug 147557 in general "When using compiz, the "Server List" window widgets do not appear, and the window cannot be closed..." [Crash,Resolved: fixed] 
[06:24] <desrt> you can build and install more of them yourself but you're basically on your own for it
[06:24] <mike> How come in Synpatic after adding the repository for something like Skype, how come and wont let me download and install it?
[06:25] <desrt> because skype is a binary for which only a 32bit version is available
[06:25] <desrt> these are all the reasons you should use 32bit :)
[06:25] <mike> but mjg59 said that 32bit apps work out of the box in 64bit.
[06:25] <desrt> mjg59; do you know if there is some way to run the CPU in 64bit mode with the exception of pointer-size?
[06:26] <mjg59> desrt: Not to the best of my knowledge
[06:26] <desrt> so you can get access to extra instructions/registers...
[06:26] <mjg59> mike: No, I said they didn't
[06:26] <desrt> mike; actually, i said that.  and it's true.  but the way debian packaging works won't allow you to do it.
[06:26] <mike> my ba
[06:26] <mike> Do you guys use 64 or 32?
[06:26] <xtknight> 32bit does work on all amd64 CPUs.  the 32bit apps dont necessarily right off, since they need to look for their 32lib libs.  sometimes "linux32 prog" will help.
[06:26] <desrt> mike; and if you wanted to install skype for yourself (outside of the packaging) you'd have to build a lot of libraries for it first
[06:27] <xtknight> otherwise put libs from i386.deb into (use "mc" to get files out of a deb) /usr/lib32/
[06:27] <desrt> i use 32bit because i don't feel like dealing with the hassles.  mjg59 uses 64bit because it's his job to fix the hassles. :)
[06:27] <desrt> (for a rather loosely defined meaning of "job", admittedly...)
[06:28] <mike> Are you guys the actually developers ofr Ubuntu?
[06:28] <desrt> some of us more than others.  me, not so much :)
[06:28] <xtknight> i'm only a user
[06:29] <mike> what about mjg59?
[06:29] <desrt> oh.  he's elite.
[06:29] <desrt> if you have a laptop, he's the reason that it works
[06:29] <mike> He is the man?
[06:29] <desrt> no
[06:29] <desrt> he's actually three men
[06:29] <mike> wow
[06:29] <desrt> ya.  seriously.
[06:29] <mneptok> and a baby
[06:29] <mike> props to you mjg59
[06:30] <mneptok> and a hydra
[06:30] <LaserJock> mike: he's on the Ubuntu Technical Board ;-)
[06:30] <LaserJock> or at least I think he still is
[06:30] <mike> Now tell me the truth do you guys use Vista or OSX?
[06:30] <mike> or true nix users?
[06:30] <LaserJock> I use OS X a lot, never even seen Vista
[06:30] <desrt> (apple doesn't release a firmware upgrader for linux...)
[06:30] <desrt> other than that i never use either
[06:31] <desrt> well, actually... i use macos when i'm trying to fix my sister's broken computer...
[06:31] <mneptok> +n
[06:32] <mike> Vista is nice eyecandy that is bloated out the wazooo
[06:32] <desrt> i've heard the eyecandy isn't even that nice :p
[06:32] <mike> better than xp then lol
[06:32] <desrt> ya... but not as good as macos from 4 years ago
[06:32] <mike> I really appreciate you guys taking the time to talk to me.
[06:33] <ScottK> IMO Windows peaked at Win2K.
[06:33] <ScottK> I will confess to mostly care about reliability, so my perspective is a bit unusual.
[06:33] <desrt> before they started changing the visual style with every release to make you think that it was new and shiny? :)
[06:33] <mike> So I guess I should but my 32bit Ubunut and reinstall over my current 64bit one?
[06:34] <desrt> mike; running a 32bit system is, without a doubt, easier
[06:34] <mike> I swear the only reason I use M$ is cause of games
[06:34] <desrt> mike; particularly if you're gonna be putting in stuff like skype
[06:34] <desrt> mike; get a wii.  i hear they're fantastic :)
[06:35] <mike> x360 for halo 3!  NO comparison
[06:35] <mike> i got a ps3 too
[06:35] <desrt> use them :)
[06:35] <mike> modded xbox
[06:35] <desrt> my aunt and her boyfriend got me to ubuntu them up a few months ago because they were sick of the crap with windows
[06:35] <desrt> his games ran like crap on ubuntu so he just stopped playing them eventually
[06:35] <desrt> then he bought a wii and is happier than ever :)
[06:36] <mike> What about adobe equivilant products.  Nice WYSIWG alternative.      Premier, Macromodia ?
[06:36] <desrt> nothing yet, unfortunately
[06:37] <desrt> but i think we'll be seeing a shift from adobe in the coming years..
[06:37] <desrt> fortunately, most "home users" don't use that stuff
[06:37] <ScottK> For PDF the tools are actually, IMO, better on Linux.
[06:37] <mike> I just like having the ability to do anything
[06:37] <desrt> ya.  evince is fantastic.
[06:37] <LaserJock> ScottK: yes, way better
[06:37] <desrt> has windows got a pdf viewer yet?
[06:37] <desrt> or do you still have to install the adobe crap?
[06:38] <mike> Windows doesnt even have a native dual pane file manager, errr
[06:38] <desrt> what happened to fileman.exe?
[06:39] <mike> Is there a WIKI for the next Gusty release?
[06:39] <mike> I would like to see what it offers
[06:39] <desrt> you can probably get a good idea by reading the release notes for the tribe releases
[06:40] <LaserJock> it's a bit tough to see, even at this point a lot can change
[06:40] <desrt> start here: http://www.ubuntu.com/testing/tribe5
[06:40] <desrt> then read backward tribe4, tribe3, ...
[06:40] <desrt> it'll give you a good overview of the new stuff that has gone in so far
[06:41] <mike> Better dual monitor support, sweet.
[06:42] <desrt> Smooth shutdown splash
[06:42] <desrt> The splash screen displayed on shutdown no longer flickers briefly to a black screen during the shutdown process. A[A[A
[06:42] <desrt> thank god :p
[06:42] <ScottK> There is also https://wiki.kubuntu.org/GutsyGibbon/Tribe5/Kubuntu to consider too.
[06:42] <mike> Do you test in VMserver are do you do a real install?
[06:42] <desrt> kubuntu is for chumps :)
[06:43] <mike> agree
[06:43] <desrt> mike; to be honest, it's probably reasonably safe at this point to install gutsy tribe on your normal machine
[06:43] <ScottK> Kubuntu is for people who actually like to have choices ;-)
[06:43] <desrt> there may be some bumps, of course.... and there will be lots of updates to download every day... but it probably works OK
[06:43] <mike> So now if I go back to 32bit, i cause I need to ditch a gb of memory or two that wont be used
[06:43] <desrt> how much ram do you have in there? :p
[06:43] <ScottK> How much memory do you have?
[06:44] <mike> Like I said I got all the power toys
[06:44] <desrt> you can run 32bit ubuntu with PAE and get access to up to about 64GB
[06:44] <mike> Was going to build another Quadcore and overclock it, but holding off the see these new 45nm CPUs coming out.
[06:44] <desrt> i doubt you have more than that :)
[06:45] <desrt> the limitation is on the amount of memory that a single process can use -- that's 3GB
[06:45] <desrt> the system as a whole can use more
[06:46] <desrt> without PAE the most memory that it makes sense to have is 4GB
[06:46] <desrt> (but a bit of it will be wasted due to memory-mapped buffers of devices)
[06:47] <kylem> no, you will get the full 4G per process on all our kernels.
[06:48] <desrt> kylem; on 32bit?  i doubt it.
[06:48] <mike> Now does linux have a MPEG AVC encoder?   FFMPEG only decodes.    And is there 1337 video editing softweare?
[06:48] <kylem> ha. whatever. don't listen to me, i don't know anything about this at all.
[06:48] <desrt> mike; probably not as per your definitions of l33t :)
[06:49] <desrt> kylem; k :)
[06:50] <mike> desrt from a surface level, not diving what is under the hood.   Gnome hands down right?
[06:50] <desrt> mike; i'm biased.
[06:50] <mike> I notice a couple KDE apps look funny i gnome.   I think I tried k3b, just looked a bit off.
[06:50] <mike> *in
[06:51] <desrt> kde and gnome use different toolkits for drawing and different themes...
[06:51] <desrt> on top of that they have rather different design philosophy and artwork
[06:51] <desrt> so it can be a bit jarring :)
[06:51] <RAOF> mike: You'd be after "x264" on the MPEG4 AVC encoding front.  You can build it into ffmpeg, if you want.
[06:53] <mike> So should I wait until october of stable Gutsy or put 7.04 32bit back on?
[06:54] <desrt> kylem; erm.... you're someone who should probably know a bit about virtual memory...
[06:54] <kylem> indeed.
[06:54] <desrt> so i don't believe you when you say you don't know anything about it
[06:55] <kylem> it was sarcasm.
[06:55] <desrt> so correct me
[06:55] <dobey> kylem: he's canadian. they don't have sarcasm there.
[06:56] <kylem> as it happens, i know a lot about virtual memory...
[06:56] <desrt> do you have some sort of a kernel-entry trampoline these days?
[06:56] <kylem> eh? we invalidate the tlb, it costs us a bit, but it's not a big deal.
[06:56] <desrt> so you _do_ have a trampoline?
[06:57] <mike> DL gutsy tribe 5 now.  Great DL speed
[06:57] <desrt> is that upstream or a patch?
[06:58] <kylem> upstream. see CONFIG_HIGHMEM4G
[06:58] <mike> Why doesnt torrent become main stream to DL large files?
[06:59] <dobey> mike: it has become maintstream
[06:59] <mike> 600 KB/s and climbing
[06:59] <LaserJock> mike: cause torrents suck :-)
[06:59] <LaserJock> at least for some people I guess
[06:59] <mike> Well I got the download page for trribe 5 and by default it sends me to http link
[06:59] <dobey> mike: gunzonline.com uses it. blizzard uses it for world of warcraft. etc...
[06:59] <dobey> it's a pain in the ass though
[07:00] <LaserJock> it always seems like torrents are soooo much slower than just using http/ftp for me
[07:01] <mike> If it is main stream then by defualt download link shouldnt link me to the torrent?
[07:02] <dobey> mike: mainstream doesn't mean everyone uses it by default
[07:02] <dobey> mainstream merely means "in widespread use"
[07:02] <minghua> Also less people downloads Tribe releases than final downloads.
[07:02] <dobey> i think numerous online games using it, pretty much classifies it as "mainstream"
[07:03] <dobey> especially with world of warcraft being one of those
[07:03] <minghua> So even if bittorrent is the default link for final release, it makes sense to go back to http/ftp for tribe.
[07:03] <mike> But does it make sense for the server?
[07:04] <dobey> the server?
[07:04] <mike> bandwith placed on server
[07:04] <LaserJock> like every time we release and the Canonical servers and mirrors get absolutely hammered :-)
[07:05] <dobey> if you're using bittorrent, there is virtually no bandwidth being used by the server for the download
[07:05] <LaserJock> yep
[07:05] <mike> bandwith cost money so I was just wondering about 'mainstream'
[07:05] <LaserJock> that's why torrents are available
[07:05] <ion_> Its a pity multicasting hasnt been more popular.
[07:05] <dobey> bandwisth ~= cost of water, at least in the US :)
[07:06] <desrt> kylem; everything i'm reading about this HIGHMEM4G tells me that it has nothing to do with virtual address space layout
[07:06] <desrt> kylem; nothing to do with the kernel/user split, anyway....
[07:06] <ion_> Id like to order twenty liters of bandwidth.
[07:06] <dobey> that'll be $0.05
[07:07] <hile> so how many bits does it fit to one liter
[07:08] <dobey> depends on what counts as a bit i guess
[07:08] <dobey> is a molecule a bit or a byte?
[07:09] <hile> a byte, I'd say
[07:09] <mike> random question.  What is the  i386 name used for 32 bit desktop edition?    I though 386 was a reference back to 1985 with the 80386 processor
[07:09] <dobey> hile: so each atom is a bit?
[07:09] <hile> not so sure
[07:09] <dobey> or are electrons bits?
[07:10] <LaserJock> the spin on the electron  is a bit
[07:12] <RAOF> Hm.  Is there any easy way to install *all* the dbgsym packages for all the packages I have installed?
[07:13] <mike> randome question #2 how come ntfs-3g made me copy the files over to ext3 partiion before being able to execute/process ?
[07:14] <kylem> huh, i thought x86 might have joined the 20th century. guess not. loss.
[07:24] <mike> About to check out tribe 5 here.   Umm.  I got a dilema.   1 blacnk HDD,   2 HDD  has two OS on it;   64 7.04 and Vista Ultimate.      I want to install Gutsy to blank hdd.    But then I would have 3 OS.   i want to get rid of 64bit 7.04
[07:26] <desrt> kylem; seems not to be the case, in fact :/
[07:26] <desrt> kylem; i like my syscalls fast anyway :)
[11:16] <Riddell> calc: congratulations
[11:22] <Amaranth> anyone know how to make apt show it's reasoning for upgrading a package?
[11:22] <Amaranth> for why it thinks an upgrade is available, i mean
[11:28] <Amaranth> i have a feeling i need mvo for that one :)
[11:38] <zasf> cjwatson: goog morning
[11:38] <geser> Amaranth: apt-cache policy pkgname
[11:38] <zasf> cjwatson: how do I know if bcm43xx-fwcutter will be in main/restricted and thus shipped with gutsy cd?
[11:38] <Amaranth> geser: no
[11:39] <Amaranth> geser: that just shows the weights
[11:39] <Amaranth> i want to know why it thinks two packages are different and one has a higher weight than the other
[11:39] <geser> ah
[11:39] <Amaranth> i've got what i'm pretty sure is a PPA bug
[11:39] <Amaranth> well, or apt
[11:40] <Amaranth> need to figure out why it always thinks the remote version is newer than the local version
[11:40] <Amaranth> no matter how many times you do the 'upgrade'
[11:40] <geser> ah, that problem
[11:40] <Amaranth> different than the original, that was fixed
[11:41] <Amaranth> this is a completely different problem that apparently only i have
[11:41] <geser> have you checked that the package with that version is only in one repo?
[11:41] <Amaranth> but it doesn't happen with the same package in the actual archive
[11:41] <Amaranth> geser: there is an older version of the package available in regular feisty
[11:41] <Amaranth> otherwise there is nothing else
[11:42] <Amaranth> apt-cache policy shows 3, 2 dupes of the new one and the old one
[11:42] <Amaranth> (one is remote the other is local)
[11:43] <Amaranth> the only thing i can think of is the ~ppa1 bit
[11:43] <Amaranth> but that doesn't do anything to the other packages i have in there
[11:45] <geser> wait for mvo or dig in the source code
[11:45] <Amaranth> i think i'll wait for mvo :)
[11:46] <Amaranth> my dig into nautilus/libeel turned me off looking at other people's code for awhile
[11:46] <stgraber> asac: ping
[12:18] <stgraber> asac: I have spent some more minutes on that open network problem with ipw3945+NM, I checked with ipw3945 without NM and the no-automatic-association (default module parameter) seems to work fine but as soon as I start the NM, it connects to my wired LAN (so it shouldn't do anything with the WLAN card) but also associate to the first open wlan it finds ...
[12:18] <stgraber> which is I'm pretty sure not a wanted behaviour at all
[12:33] <Treenaks> same with ipw2200 -- that associates to the first AP it can the moment you load the driver
[12:34] <Treenaks> (which might be illegal in some places...)
[12:34] <Amaranth> Treenaks: dude, that is no annoying
[12:35] <Treenaks> Amaranth: 'no' -> 'so' or 'not'?
[12:35] <Amaranth> so :)
[12:36] <Amaranth> before i login i'm on the neighbor's 'linksys' network
[12:36] <Amaranth> that is illegal here :)
[12:37] <Treenaks> Amaranth: here it's only illegal if there's a "no access" sign (meaning "it's encrypted" or "the ssid contains words that can be interpreted as such")
[12:37] <Treenaks> "linksys" and "default" are the new "Welcome" signs :)
[12:39] <Amaranth> hehe
[01:14] <Kopfgeldjaeger> hi
[03:26] <jc-denton> hi all
[03:27] <jc-denton> in never ubuntu distros disks are always called /dev/sdx
[03:27] <jc-denton> is that correct
[03:27] <jc-denton> how can i figure out if the disk behind is ata, sata or scsi
[03:27] <jc-denton> because i can't enable 32bit io and dma so far
[04:16] <Nafallo> !weekend
[04:16] <ubotu> It's a weekend.  Often on weekends, the paid developers, and a lot of the community, may not be around to answer your question.  Please be patient, wait longer than you normally would, or try again during the working week.
[04:21] <pkern> jc-denton: hdparm -I <dev>
[04:21] <pkern> jc-denton: That won't tell you the type of disk, but the capabilities.
[04:22] <pkern> jc-denton: It will also show you which DMA mode is currently used (asterisk). But I don't think that you have to mess with that nowadays.
[05:26] <cjwatson> whoopsie. sorry for breaking hibernate for everyone. fixed in powermanagement-interface 0.3.17
[05:35] <whiprush_> meffie: hi
[06:01] <IntuitiveNipple> cjwatson: Thanks for that :)
[06:37] <DoctorMO> prod: https://wiki.ubuntu.com/PimSyncPlan
[06:37] <DoctorMO> I'd like to get feedback on my ideas, something other than 'wow you'd be my hero'
[06:40] <pygi> DoctorMO, :P
[06:43] <LaserJock> yes
[06:43] <LaserJock> we are full of good ideas that never get implemented :-)
[06:44] <bluefoxicy> so it looks like Gutsy has totally destroyed the separation of privileges security model on ssh :D
[06:44] <bluefoxicy> is anyone here running feisty still that can snag me something?
[06:45] <LaserJock> bluefoxicy: depends on what it is ;-)
[06:45] <bluefoxicy> LaserJock:  find a pid of sshd, cat /proc/$pid/status | grep CapEff
[06:45] <bluefoxicy> actually, just grep Cap
[06:46] <LaserJock> CapInh: 0000000000000000
[06:46] <LaserJock> CapPrm: 00000000fffffeff
[06:46] <LaserJock> CapEff: 00000000fffffeff
[06:46] <bluefoxicy> hmm.
[06:47] <bluefoxicy> LaserJock:  check dmesg and see if you see something like "Capability LSM initialized as secondary" twice, one late in the boot process (30 seconds or so)
[06:47] <bluefoxicy> this may not be a gutsy thing (I'm thinking it's the way AppArmor is done right now, but perhaps not...)
[06:48] <IntuitiveNipple> I only see the one in Fiesty 32-bit "[    2.467828]  Capability LSM initialized"
[06:48] <LaserJock> [   20.021888]  Capability LSM initialized
[06:48] <LaserJock> same here, only 1
[06:48] <bluefoxicy> LaserJock:  that's probably way too late.
[06:49] <bluefoxicy> when caps are initialized that late in the boot process, any process before them gets full caps (i.e. the whole separation of privileges model used by things like ssh or anything else that drops caps is destroyed)
[06:49] <bluefoxicy> IntuitiveNipple:  can YOU check to see the CapEff on sshd?  Yours is probably running under something less permissive, I'm guessing...
[06:50] <IntuitiveNipple> bluefoxicy: The same as LaserJock's
[06:50] <IntuitiveNipple> CapInh: 0000000000000000
[06:50] <IntuitiveNipple> CapPrm: 00000000fffffeff
[06:50] <IntuitiveNipple> CapEff: 00000000fffffeff
[06:51] <bluefoxicy> hmm.  I'm getting the same string after restarting ssh.  Maybe I'm wrong on how this works.
[06:53] <bluefoxicy> LJ, IN:  thanks.
[06:53] <LaserJock> np
[06:55] <cjwatson> bluefoxicy: ssh doesn't use capabilities as such, just seteuid etc.
[06:58] <cjwatson> bluefoxicy: I suspect you'll find the sshd network child process still drops its privileges just fine
[06:58] <bluefoxicy> cjwatson: hmm.  Theo gave a presentation touting the capabilities sep.. oh ok
[06:58] <bluefoxicy> that's probably working then, since the child process will exec() angd get caps.
[06:59] <cjwatson> bluefoxicy: you misunderstood the word "capabilities". also OpenSSH people always talk about "privilege separation" not "capability separation".
[06:59] <bluefoxicy> cjwatson:  they talk about privilege separation using capabilites IIRC
[06:59] <cjwatson> wrong.
[07:00] <cjwatson> sshd forks a child which chroots to an empty jail, does setgroups+setregid+setreuid, listens to network communication, and passes sanitised versions of that communication over a simple protocol to the parent.
[07:00] <cjwatson> so parsing bugs can do much less damage
[07:01] <cjwatson> that's what sshd privsep means
[07:01] <bluefoxicy> hmm o.o
[07:01] <bluefoxicy> I need to research this more.  That seems sane though.
[07:02] <cjwatson> oh and that process is only used during authentication, so you won't see it hanging around afterwards.
[07:03] <cjwatson> i.e. no amount of grep Cap /proc/*/status will show anything interesting unless you nobble sshd to hang around during authentication a bit longer
[07:05] <cjwatson> (after auth, the child with user privileges is perfectly good for parsing network data)
[07:07] <zul> heylo
[07:07] <ion_> hiho
[07:12] <ion_> Really nice that xubuntu-desktop switched to gnome-screensaver.
[07:16] <DoctorMO> LaserJock: yea
[07:46] <jdong> https://wiki.ubuntu.com/DevelopmentCodeNames
[07:46] <jdong> "Hard Heron with 5 years of support!"
[07:47] <IntuitiveNipple> You just wait for Grumpy!
[07:49] <IntuitiveNipple> is that a flavour of hog? :p
[07:53] <IntuitiveNipple> For "I", I'd love to see 'Inquisitive Imp' :)
[08:24] <alesan> re
[08:26] <alesan> do you have any idea what is the package for the "logout" button? gnome people told me it is ubuntu-patched
[08:33] <LaserJock> alesan: boy, I have no idea
[08:35] <alesan> any idea how to find out? I would very much like to disable the "fading" thing
[08:36] <ScottK> If I were to guess it would be in gdm, but no promises.
[08:40] <alesan> ok thanks
[08:40] <alesan> bye
[08:43] <gnomefreak> ski would think either gdm or gnome-session
[08:43] <gnomefreak> gnome-session-manager if not gnome-session
[09:17] <mike> desrt you here?
[09:22] <mike> desrt told me to use 7.10 tribe 5, but my monitor res is acting up big time
[10:56] <asac> stgraber: there?
[10:59] <stgraber> asac: yep
[10:59] <asac> cool
[10:59] <asac> sorry been out for the day ;)
[11:00] <asac> stgraber: are you still in the same environment (where nm associates with the open-network?)
[11:00] <stgraber> yes
[11:01] <stgraber> any idea why NM would associate with the first network it finds even if it's connected to wired and then shouldn't do anything with vlan ?
[11:01] <asac> good what does sudo iwlist scan give you?
[11:01] <asac> stgraber: does network-manager know about the first network?
[11:02] <asac> stgraber: e.g. is it displayed as connected ... or did you find out with ifconfig et al
[11:03] <stgraber> it's not marked as connected but I can see the associate using iwconfig
[11:03] <asac> yes ok that makes sense then
[11:04] <stgraber> my current test is : connect to wired LAN, then kill ipw3945d and unload ipw3945, then reload it
[11:04] <asac> stgraber: can you explicitly switch back to wired after that?
[11:04] <stgraber> NM is still connected to the wired LAN but iwconfig shows the association to the open net
[11:04] <mjg59> asac: I can't associate my ipw3945 with an open network. It seems the driver never sends the association event.
[11:04] <asac> mjg59: yes for open networks its probably a driver bug
[11:04] <stgraber> doing the same with NM turned off doesn't automatically associate
[11:05] <asac> mjg59: however it appears to work better with wpa supplicant ... no idea why its different
[11:06] <stgraber> asac: what do you mean by "after that", at this point I never disconnect from wired LAN nor asked NM to connect to a WLAN
[11:17] <asac> stgraber: so you are just associated without ip?
[11:17] <stgraber> yes
[11:18] <asac> stgraber: ok then nm didn't receive the association event
[11:18] <stgraber> I'm connected on my wired network, then load the ipw3945 driver and a few seconds after iwconfig shows an association
[11:18] <stgraber> but I never asked NM to connect to a wifi network :)
[11:18] <asac> stgraber: what do you see in logs?
[11:19] <asac> does nm officially try to bring up the interface?
[11:19] <stgraber> no
[11:20] <stgraber> http://paste.stgraber.org/3263
[11:20] <asac> the logs don't even mention the device?
[11:20] <stgraber> ^ What happens after the "modprobe ipw3945"
[11:21] <asac> stgraber: ok ... can you try to explicitly reconnect to the wired network
[11:21] <asac> then do the driver reload?
[11:23] <stgraber> exactly the same behaviour, I load the driver, the second after it's connect to an open wlan
[11:38] <deadchip> is Scott Tankard around maybe?
[11:39] <deadchip> i've got something for the automated-testing proposals
[11:39] <deadchip> i'm the upstream maintainer of BMP, and the old XMMS based BMP is proposed for the automated testing stuff, but it's an deprecated app
[11:39] <deadchip> BMPx has replaced it since a good while from our side, and it would be much better if it would receive the testing instead
[11:40] <deadchip> it's available in Debian and thus to ubuntu users at the moment through Debian
[11:40] <deadchip> however I don't want to just change this in the wiki
[11:40] <deadchip> anyone knows whom i can contact about this?
[11:41] <deadchip> ermm Scott T. is just the last editor of the page i don't know if he's the guy responsible for this
[11:44] <ScottK> A link to the page you are discussing wouldn't hurt.
[11:44] <mhb> yes
[11:44] <mhb> ScottK: I thought the same thing :o)
[11:45] <asac> stgraber:
[11:45] <asac>         /* Device must be up before we can scan */
[11:45] <asac>         if (nm_device_bring_up_wait (NM_DEVICE (self), 1))
[11:47] <stgraber> asac: yes, but doing : modprobe ipw3945 + ifconfig eth1 up doesn't associate without NM
[11:47] <crimsun> deadchip: to which URL are you referring?
[11:47] <deadchip> crimsun: https://wiki.ubuntu.com/Testing/Suite/Desktop
[11:48] <deadchip> well i've changed it now, as the edit history showed other people added their apps there, etc
[11:48] <deadchip> it's really no use at all to add BMP to the testing because upstream (we) won't be making any changes to it
[11:48] <crimsun> right, that's fine.
[11:48] <geser> doko: you forgot to change the maintainer field for your last upload of gcc-snapshot to Debian unstable
[11:48] <deadchip> so everything that would be a potential fix would end with ubuntu developers
[11:48] <deadchip> ok
[11:54] <asac> stgraber: what network manager does from what i see is just unsetting the essid to "" ... then scanning
[11:54] <asac> maybe you get an association if you just do that?
[11:56] <stgraber> asac: ok, just tried and indeed doing : iwconfig eth1 essid "" makes the card to associate
[11:57] <asac> ok then its either a feature or a driver bug
[11:57] <asac> so what issue is left?
[11:57] <asac> you cannot connect to open networks right?
[11:59] <stgraber> asac: yes
[11:59] <asac> ok
[11:59] <asac> you have nm sources?
[12:00] <stgraber> asac: and I think that automatic association might be the problem (as it shouldn't be associated and asking NM to connect to a public WLAN doesn't make it associate to the selected one)
[12:00] <asac> stgraber: nm-device-802-11-wireless.c line 531 533
[12:00] <stgraber> I think so but I'll have to update them anyway (didn't pull for a while)
[12:00] <stgraber> asac: ubuntu.0.6.x ?
[12:00] <asac> there is are association / disassociation event cases there that appear to be not processed
[12:01] <asac> stgraber: yes
[12:01] <asac> stgraber: maybe add nm_info("wireless_event_helper - associated event");
[12:02] <asac> stgraber: and nm_info("wireless_event_helper - disassociated event");
[12:02] <asac> to see if you get those events
[12:02] <stgraber> k
[12:04] <asac> stgraber: while you build ... can you post a log of a connect attempt to open network?
[12:05] <stgraber> I'm updating my pbuilder, as soon as I no longer need the network I'll
[12:05] <asac> k
[12:08] <stgraber> asac: http://paste.stgraber.org/3288
[12:09] <asac> stgraber: what does iwconfig show you?
[12:09] <asac> stgraber: are you associated?
[12:09] <asac> when it hangs in stage 1
[12:10] <stgraber> yes I'm asociated but I also was before asking NM to connect
[12:11] <stgraber> s/also/already/
[12:11] <asac> yes right
[12:15] <asac> stgraber: its starnge that you only see "DevicePrepare ) scheduled" ... but not *started*
[12:15] <stgraber> asac: http://paste.stgraber.org/3289
[12:15] <asac> its not much that could go wrong ... its basically just an idle timeout
[12:15] <stgraber> with the two nm_info lines
[12:16] <asac> stgraber: ok can you do a connect attempt now
[12:16] <asac> and wait till nm gives up ... at best note which time that is in log :)
[12:17] <stgraber> I doubt it'll give up, I always had to kill it as I wasn't even able to select wired connection :)
[12:17] <asac> huh?
[12:17] <asac> hmm
[12:17] <asac> sounds reasonable
[12:17] <asac> maybe a dead log?
[12:18] <asac> i mean i wondered why preparation is not started for you ... its basically just attaching and idle source
[12:18] <asac> so not much that could go wrong
[12:18] <asac> s/dead log/deadlock/
[12:21] <stgraber> this try seems a bit better (which is strange as nothing is different except the two nm_info ...) : http://paste.stgraber.org/3290
[12:21] <stgraber> and I was able to switch back to wired LAN
[12:21] <stgraber> but no disassociated/associated event line shown :(
[12:22] <asac> stgraber: i don't see any association event now
[12:22] <asac> didn't you reload driver?
[12:25] <stgraber> this log start just before I click on the open wlan name in NM, the NM auto-association was before I think
[12:25] <asac> but you are associated to another network? or the right one?
[12:25] <stgraber> I only have one open network here, so it's the same network I'm automatically associated to and I'm trying to connect
[12:26] <asac> stgraber: ok now switching rflink ?
[12:26] <asac> what happens then?
[12:27] <stgraber> switching rflink ?
[12:27] <asac> well turn radio off and on
[12:27] <asac> there should be something :)
[12:27] <stgraber> a ok, let me try
[12:30] <stgraber> works
[12:30] <asac> what did yousee in log?
[12:30] <asac> did you just get an assocaited event?
[12:31] <stgraber> if I do it while it's at stage 1 or just before it doesn't work (iwconfig shows unassociated), if it's at stage 2 or after it just works fine