[00:00] <SpamapS> slangasek: thanks! :) I had a much simpler one that basically just does "Waiting for network configuration ..." .. I suppose thats the right thing to do for a first pass. :)
[00:01] <slangasek> SpamapS: "if there's only an 'auto ethX' line" - if that's the only line, with no iface, /etc/init/network-interface.conf certainly won't bring it up
[00:01] <seq> cjwatson: Problem: About 40 minutes ago I removed the 'auto eth0' line
[00:01] <cjwatson> then put it back from memory and pastebin that :-)
[00:02] <SpamapS> slangasek: you sure?
[00:02] <slangasek> SpamapS: er, absolutely
[00:02]  * SpamapS just had it boot like lightning in that configuration
[00:03] <slangasek> SpamapS: if the only thing in /etc/network/interfaces is the 'auto' line, there's nothing that tells ifup to do anything with the interface, only that it's *allowed* to be brought up automatically
[00:03] <SpamapS> ok, this is weird
[00:03] <slangasek> so it won't be brought up by ifupdown, ever
[00:03] <SpamapS> it actually dhcliented it
[00:04] <slangasek> not NM?
[00:04] <SpamapS> http://paste.ubuntu.com/688722/
[00:04] <SpamapS> no NM installed
[00:04] <slangasek> -sf /usr/lib/NetworkManager/nm-dhcp-client.action
[00:04] <slangasek> really? ;)
[00:04] <SpamapS> oh wait
[00:04] <SpamapS> HAHAH I did install n-m on this to test something
[00:05] <SpamapS> argh
[00:05] <seq> cjwatson: /etc/network/interfaces: http://paste.ubuntu.com/688725/
[00:06] <seq> cjwatson: /var/log/installer/syslog: http://paste.ubuntu.com/688724/
[00:06] <seq> cjwatson: I actually found the blank media, it was an amd64 alternate daily from 20110827
[00:07] <SpamapS> ok so yeah, if n-m is installed, and there's only an auto line, the system ends up booting because n-m starts and brings it up
[00:08] <SpamapS> (and its plugged in, of course)
[00:09] <seq> SpamapS: /run/network http://paste.ubuntu.com/688728/
[00:09] <seq> SpamapS: note that it isn't my boot state anymore
[00:09] <SpamapS> seq: is that after a 2 minute delay?
[00:09] <seq> I've brought stuff up since then
[00:09] <SpamapS> seq: yeah thats not as helpful, but I think I get whats going on now anyway. Thanks.
[00:10] <seq> no. I'd have to reboot. Unfortunately I only have the one machine currently, so I can't run my stable natty at the same time as my testing oneiric :/
[00:10] <slangasek> SpamapS: right; so this is a bug in /etc/network/if-up.d/upstart:get_auto_interfaces(), treating things as auto interfaces that aren't really interfaces at all from ifupdown's perspective
[00:12] <slangasek> probably needs to use ifquery + the check for auto
[00:13] <cjwatson> seq: ok, so there was an iface line, but /usr/lib/NetworkManager/ifblacklist_migrate.sh removed it
[00:13] <SpamapS> slangasek: ifquery has some kind of braindead notion of aliases that made it unsuitable IIRC
[00:13] <SpamapS> slangasek: I'll ask smoser about it tomorrow.. I'm pretty much burnt out on debugging it for today.
[00:13] <cjwatson> so therefore the upstart job needs to consider auto-but-no-iface as "ignore, it's NM's problem"
[00:13] <cjwatson> since that's what the NM blacklisting script does
[00:14] <seq> I'm free to test if I'm waiting on network with the original /etc/network/interfaces file with SpamapS' failsafe.conf from his proposed branch above. Any other things I should be aware of or consider?
[00:14] <SpamapS> I have to agree though, that we should filter it out if it has no config in /e/n/i
[00:14] <seq> or think you've got a handle on it?
[00:14]  * cjwatson shoves back in the "not an installer problem" bin :-)
[00:14] <slangasek> SpamapS: right, I'll be interested in the details as I was pretty sure I accounted for aliases when coding up ifquery
[00:15] <SpamapS> slangasek: I believe the problem was that one couldn't easily match them up to the auto lines.. but my memory is fuzzy.
[00:15] <SpamapS> cjwatson: thanks for looking at this so carefully. :)
[00:16] <SpamapS> whole brain is fuzzy actually.. I'll get back on this in the morning.
[00:16] <smoser> slangasek, you are sure that it is ?
[00:16] <smoser> that it is a bug in that code identifying things as auto that are not ?
[00:16] <slangasek> smoser: there is definitely a bug in the code, yes
[00:16] <SpamapS> smoser: the code just looks for auto lines, it doesn't verify that they have configuration
[00:16] <slangasek> it waits for 'auto' interfaces that are never defined
[00:17] <smoser> so what should it be looking for ?
[00:17] <slangasek> it needs to check both for the auto line and ask ifupdown to confirm that there's actually an interface definition
[00:18] <smoser> ask ifupdown
[00:18] <slangasek> yes, i.e. ifquery
[00:18] <slangasek> to avoid writing your own interfaces(5) parser
[00:19] <cjwatson> and since ifquery is an Ubuntu extension to begin with, it makes no sense to work around it rather than fixing it if it's broken
[00:19] <smoser> agreed.
[00:21] <smoser> the parser i have should support aliases:q
[00:21] <seq> One more unrelated question for you: What does Meta+p do? In natty+unity I had this bound to move a window to top/right, but in oneiric+unity it just flashes the screen and causes X to hang for about five seconds.
[00:21] <SpamapS> smoser: you mean mappings right?
[00:21] <SpamapS> smoser: I'm trying hard to remember why it wasn't suitable
[00:23] <smoser> when we were using ifstate it had sutff that worked to decide if something was up
[00:23] <smoser> it read ifstate
[00:23] <smoser> but now we dont read ifstate
[00:23] <SpamapS> AH
[00:25] <SpamapS> smoser: so we can just change get_auto_interfaces to use ifquery ..
[00:25] <smoser> slangasek, do you have a suggestion on how to call ifquery ? ifquery --no-mappings ?
[00:25] <smoser> i dont know. maybe. that would be nice
[00:26] <SpamapS> smoser: as long as thats the interface name that comes in as $IFACE
[00:27] <slangasek> smoser: not sure yet, but I guess 'ifquery --allow auto --list --no-mappings' is how we would /want/ it to work
[00:28] <slangasek> (the --no-mappings doesn't seem to work so well in practice at the moment)
[00:28] <smoser> it seems you dont need --allow auto
[00:28] <smoser> :)
[00:29] <smoser> i just tested that if i commented out 'auto eth0' and 'ifquery --list --no-mappings' then did not show it.
[00:29] <SpamapS> smoser: I think thats just the default
[00:30] <SpamapS> what I'm confused by is how to test the mappings
[00:30] <smoser> and i'm not sure that --no-mappings is functioning right.
[00:30] <smoser> well. anyway.
[00:30] <smoser> we can come up with a solution here.
[00:31] <YokoZar> slangasek: What's the status of OpenSSL 32 bit these days?
[00:32] <SpamapS> smoser: yeah, --no-mappings is busted
[00:32] <YokoZar> slangasek: The current build log for Wine:  https://launchpadlibrarian.net/79822996/buildlog_ubuntu-oneiric-amd64.wine1.3_1.3.28-0ubuntu1_BUILDING.txt.gz  (ctrl+F for not found)
[00:32] <SpamapS> anyway, brain fried, and time for karate
[00:33] <smoser> at very least 'ifquery --list 2>/dev/null' seems to ignore 'auto' interfaces that do not have configured section
[00:34] <SpamapS> smoser: it shows mapped interfaces though, whereas the real interface seems to be passed to the scripts
[00:34] <smoser> SpamapS, do you have an example interfaces ?
[00:35] <SpamapS> smoser: auto eth0-foo\niface eth0-foo inet dhcp\n
[00:35] <SpamapS> smoser: sudo ifup eth0=eth0-foo ...
[00:35] <SpamapS> smoser: eth0 is passed as $IFACE
[00:35] <SpamapS> smoser: but ifquery -a --list shows eth0-foo only
[00:36] <smoser> is that a valid config ?
[00:36] <SpamapS> smoser: yep, brought up the physical eth0 as the logical eth0-foo
[00:37] <SpamapS> smoser: tho.. we can use $LOGICAL to check
[00:37] <smoser> SpamapS, you should go to karate
[00:37] <SpamapS> Yeah, late == pushups :)
[00:37] <smoser> but that is not a config we're worried about i think
[00:37] <SpamapS> smoser: I think $LOGICAL is the way to go
[00:37]  * SpamapS is gone
[00:38] <smoser> because in that config, eth0-foo is not an auto interface. and it wont be brought up by ifup on first boot.
[00:38] <smoser> yeah
[00:39] <smoser> i think its fine'
[00:49] <slangasek> YokoZar: libssl0.9.8 is still included in ia32-libs; though perhaps the fact that I was hand-hacking it into the source package caused the dev symlink to not be where it should and we need to regenerate ia32-libs now that openssl0.9.8 source is in the archive?
[00:55] <YokoZar> slangasek: I'll experiment with an ia32-libs freshening, see if that clears it
[01:00] <slangasek> SpamapS: 'sudo ifup eth0=eth0-foo' - I don't think that's how it would be invoked in practice though
[01:57] <YokoZar> slangasek: btw I'm gonna try building true multiarch wine and see if it's less broken ;)
[01:58] <YokoZar> (it will likely build at this point, however it will complain about 15 different things being disabled since they're not multiarched yet)
[03:08] <Seq> Is there a way I can force myself to end up at a basic shell in my initrd before my disks mount?
[03:50] <pitti> Good morning
[03:51] <pitti> bdmurray: hm, I don't see anything wrong with the cronjob
[03:51] <stgraber> Good morning pitti
[03:52] <stgraber> still wondering how you manage to wake up and start working before 6am ;)
[03:54] <pitti> stgraber: I actually had thought it was 6 am, but seems my wife got up earlier today
[03:55] <pitti> *yawn*
[03:56] <stgraber> pitti: hehe :)
[04:21]  * TheMuso is a 6am riser, although I don't jump on the computer for a good 2 hours after I get up. :)_
[06:58] <dholbach> good morning
[07:12] <doko> debfx, ScottK: please could you address this within the kubuntu umbrella: bug 770789
[07:46] <doko> Daviey, please have a look at bug 756107
[07:53] <doko> debfx, ScottK: same for bug 832856
[08:21] <doko> pitti: please have a look at bug 832778 (pg 9.1 issue)
[08:22] <pitti> doko: thanks, will do
[08:43] <doko> jelmer, please could you have a look at bug 832901?
[09:04] <jelmer_> doko, ok
[10:03] <alexbligh1> Is it too late to request an update of Oneiric for a minor point release from upstream? (specifically bird-1.3.2 to bird-1.3.3 and ditto bird6). If not, how do I do it?
[10:06] <dupondje> æ­£å¨å¤çç¨äº gnome-menus çè§¦åå¨...
[10:07] <dupondje> suddenly there are things in Chinese ?! :)
[10:28] <Daviey> I think i am doing it wrong.
[10:30] <Laney> alexbligh1: if they are bug fix releases, then that's easy (request a sync). otherwise, it's a bit harder but can still be done if necessary (request a freeze exception)
[10:30] <Laney> !sync
[10:30] <Laney> !ffe
[10:39] <Daviey> cjwatson: I've created a udeb for libxmlrpc, but it also seems i need one for libcurl, and now libidn.. am i doing it wrong?
[10:39] <cjwatson> if you're using libraries, you need them to be in udebs
[10:40] <cjwatson> libcurl sounds an extremely unfun thing to drag in what with the openssl/gnutls duality
[10:40] <Daviey> cjwatson: Well libcurl3.so seemed to be enough to satisfy the concern,
[10:41] <Daviey> I'm trying to work out why the heck it needs libidn
[10:46] <cjwatson> I do have to question whether it's still appropriate to be adding udebs at this stage ...
[10:46] <cjwatson> especially if you're running into trouble
[10:47] <Daviey> cjwatson: Well yes, that is a valid concern.
[10:49] <alexbligh1> Laney, thanks. It's mostly bugfixes, including failing to use /etc/default/bird (which is my problem)
[11:13] <ScottK> doko: 832856 is now a removal bug.
[11:23] <ScottK> doko: 770789 too.
[12:05] <sastudio> Hello,is there a way to get the last path that Nautilus accessed?
[12:05] <sastudio> from a bash script or C
[12:10] <sastudio> Anyone?
[12:22] <ScottK> sastudio: That sounds like a question for #ubuntu.  This channel is for development of Ubuntu, not support.
[12:23] <ScottK> cjwatson: Thanks for taking a crack at the boost related failures.  I'll see if there are any in the other archs stack I can figure out.
[12:25] <cjwatson> ScottK: I think I'm out of steam for now, but have finished frogatto
[12:25] <ScottK> Cool.
[12:26] <sastudio> Forgive my mistake,since I needed it for developing a little project for Ubuntu I thought this was the right one.So no external ubuntu projects in here?
[12:30] <didrocks> sastudio: (see /topic): #ubuntu-app-devel for application development on Ubuntu
[12:31] <sastudio> @didrocks Thank you,Sir!
[12:31] <udevbot> Error: "didrocks" is not a valid command.
[12:31] <desrt> sastudio: may also want to try #nautilus on irc.gnome.org, but i suspect the answer to your question is 'no'
[12:33] <micahg> slangasek: is there any was to use a bigger hammer for people on oneiric to enable multiarch?  we keep getting bug reports about flash not being installable
[12:34] <slangasek> micahg: infinity was going to add a config file to dpkg
[12:34] <slangasek> which would help us sweep up the people who run alphas but don't read release notes or u-d-a :)
[12:36] <micahg> slangasek: that sounds great, thanks
[13:08] <pitti> slangasek: hm, flicker-free boot doesn't work for me with gdm either
[13:08] <pitti> slangasek: but anyway, great that you got this fixed!
[13:09] <slangasek> pitti: define "not work"?
[13:09] <pitti> (will take a look at the FFE soon, just saw it coming in)
[13:09] <pitti> slangasek: it looks exactly like lightdm -- plymouth, then text, then X/gdm
[13:09] <slangasek> hmm
[13:09] <pitti> I thought it wasn't DM specific
[13:09] <slangasek> I've not seen that at all
[13:09] <slangasek> could be that something kills plymouth early for you for some reason
[13:11] <slangasek> pitti: anyway, the flicker-free needs wider hardware testing before I'm comfortable landing it post-beta... any chance you have !intel video? :)
[13:11] <pitti> slangasek: all intel here ..
[13:11] <slangasek> ok
[13:11] <slangasek> maybe you at least have a non-cryptsetup system, to test that case?
[13:11] <pitti> I guess a more widespread call for testing might trigger more feedback?
[13:11] <pitti> slangasek: yes, I have switched to ecryptfs a year or two ago
[13:12] <slangasek> yay
[13:12] <pitti> slangasek: but I know some people in my team with ati/nouveau :)
[13:13] <pitti> oh, http://bazaar.launchpad.net/~vorlon/ubuntu/oneiric/plymouth/lightdm-integration/revision/1398 is all that's needed?
[13:13] <pitti> i. e. that can easily be applied manually on test systems, no need to enable a PPA etc.
[13:13] <slangasek> would be great to have their testing :)  if we need to put out a wider call, we can do that too
[13:13] <slangasek> yes, it's a small change
[13:13] <slangasek> just make plymouth-stop not stop
[13:22] <pitti> slangasek: heh, without cryptsetup I only see plymouth for some 0.3 seconds :)
[13:22] <slangasek> yep, that's pretty typical
[13:22] <pitti> and a blinking text cursor for like 10 s before
[13:23] <pitti> for some reason, looking for the root fs takes ages these days (http://people.canonical.com/~pitti/bootcharts/donald-oneiric-20110902.png)
[13:23] <slangasek> the problem is it's a tradeoff between bringing up the splash early, or booting faster
[13:23] <slangasek> hmm
[13:24] <pitti> I think that's the main reason why I install cryptsetup these days :)
[13:24] <pitti> (well, also because I still have an encrypted USB stick)
[13:24] <slangasek> the wait-for-root is within resume.  do you have a broken resume setting somewhere on the disk?
[13:25] <pitti> hm, this is a fresh beta-1 install
[13:25] <pitti> but I kept the original partitioning
[13:25] <pitti> so maybe the swap partition still has some cruft
[13:26] <pitti> $ cat /etc/initramfs-tools/conf.d/resume
[13:26] <pitti> RESUME=UUID=46fc4d23-f197-4536-b15c-99ac162fa2df
[13:26] <directhex> lamont, would it be possible to arrange a non-package build on one of the i386 builders, since the only machines in the universe which reliably reproduce the problem we're showing are vernadsky and one other of the ubuntu buildds? i don't want to break the published package in oneiric with a speculative patch
[13:26] <pitti> that one?
[13:26] <pitti> slangasek: oh, http://paste.ubuntu.com/689186/ - might be that cryptsetup did that
[13:27] <lamont> directhex: ew
[13:28] <Laney> heh
[13:28] <directhex> lamont, ew indeed
[13:28] <lamont> directhex: give me a sourcepackage in a ppa somewhere
[13:28] <directhex> lamont, you're a star :)
[13:29] <slangasek> pitti: I don't think cryptsetup touches /etc/initramfs-tools/conf.d/resume
[13:29] <pitti> slangasek: then it's probably ubiquity when you enable/have ecryptfs
[13:29] <slangasek> could be
[13:32] <micahg> lamont: I got a chroot error on an armel panda buildd, do you need anything from it or should I retry?
[13:32] <lamont> micahg: which machine?
[13:32] <micahg> lamont: nihal
[13:32] <lamont> build record?
[13:33] <micahg> lamont: https://launchpad.net/ubuntu/+source/gimp/2.6.11-2ubuntu3/+build/2784357
[13:36] <lamont> micahg: I tossed it back into the pool just to see what it has to say
[13:36] <lamont> the exact failure bothers me.
[13:36] <lamont> data integrity issues are always an interesting corner case
[13:36] <kirkland> pitti: slangasek: what's the ecryptfs question?
[13:37] <micahg> lamont: ok, thanks
[13:37] <lamont> and cleaned out filecache-default on the box, I'll go give back all the chrootwaits in a moment
[13:39] <slangasek> kirkland: the question is, what created the bogus /etc/initramfs-tools/conf.d/resume on pitti's system that slows down the initramfs waiting for a device that apparently isn't there
[13:39] <cjwatson> some bit of the installer creates that
[13:39] <cjwatson> if the UUID then changes, probably nothing updates it
[13:40] <cjwatson> base-installer and ubiquity both do it
[13:41] <cjwatson> AFAIK the installer has to do that or else hibernation can't work
[13:41] <cjwatson> it has nothing to do with ecryptfs anyway
[13:41] <pitti> kirkland, slangasek: confirmed, when I clean up the conf.d/resume bit, it's 5 secs faster
[13:42] <pitti> does hibernation even work on an encrypted swap?
[13:42] <kirkland> pitti: slangasek: yeah, i concur with cjwatson;  I don't know of any part of ecryptfs that would have touched that
[13:43] <pitti> kirkland: not ecryptfs itself; I mean, ubiquity probably sets up an encrypted swap if you choose to encrypt your home dir
[13:44] <kirkland> pitti: yes, it used to, anyway, using ecryptfs-setup-swap
[14:59] <directhex> lamont,  https://launchpad.net/~directhex/+archive/monoxide/+sourcepub/1940712/+listing-archive-extra
[14:59] <directhex> lamont, i assumed you'd be happier if i gave it time to test-build successfully before you waste your time with it
[14:59] <lamont> heh. ta
[15:09] <directhex> lamont, previously it was reliably failing to build when building the docs (look for cs-errors in the build log) on vernadsky and... i think rothera. definitely vernadsky. upstream's patch should cause it to build okay but print some debug spew when it would previously fail (the debug spew would be helpful to pass to upstream)
[15:10]  * lamont stuffs vernadsky on manual
[15:12] <lamont> directhex: and this is mono, yes?
[15:12] <directhex> lamont, yes
[15:12] <lamont> 2.10.5-1+wtf1 to be specific
[15:14] <directhex> lamont, yes
[15:15]  * lamont puts all of i386/archive on manual, waits for zirconium to finish failing the build he killed
[15:18] <lamont> directhex: there, building
[15:18] <lamont> on vernadsky, even
[15:18] <directhex> lamont, brilliant
[15:18] <lamont> directhex: can you see the build?
[15:18] <lamont> https://launchpad.net/~lamont/+archive/non-virt/+build/2784832
[15:18] <directhex> https://launchpad.net/~lamont/+archive/non-virt/+build/2784832
[15:18] <directhex> hah, yes then ;)
[15:19] <directhex> lamont, and this way there's even a proper log. you're a star
[15:19] <lamont> anyway, I'm going to leave it in your lap then
[15:19] <lamont> but I'm going to kill the armel build
[15:19] <lamont> unless there's some reason to leave it running
[15:20] <lamont> it might be nice to see how long a panda takes to build mono
[15:21] <micahg> +1, I'd be interested in seeing how long mono takes on a panda
[15:22] <lamont> micahg: I shall let it live then
[15:22] <directhex> longer than it used to. building the (C) runtime 4 times takes a while. the c# compiler is pretty fast on arm
[15:23] <micahg> directhex: well, we can see apples to apples, that build took 7 hours on the old armel machines
[15:23] <directhex> okay
[15:25] <lamont> micahg: bbg3 or beaglexm, I wonder
[15:34] <doko> hmm, why doesn't libjack-dev depend on jack1d?
[15:34] <doko> -dev has a .so symlink to a library in jack1d
[15:40] <doko> I'll call adding libraries in LDFLAGS monkey linking ...
[15:45] <cjwatson> ScottK: uploaded libnet-server-perl from pkg-perl - testing in an amavis context would be welcome since this is entirely different from the patch the amavis people suggested
[15:45] <ScottK> cjwatson: Thanks.  I'll see if I can find someone to test.
[15:51] <cjwatson> jelmer: are you aware of bug 831374?
[15:59] <Laney> directhex: we have point passingness
[16:00] <directhex> Laney, good good - but i missed the failure point, so need to wait for completion to access the log
[16:00] <Laney> it didn't say anything
[16:00] <Laney> maybe earlier
[16:00] <Laney> did you export MONO_DEBUG?
[16:01] <directhex> ..........
[16:01] <directhex> sigh. this is why i shouldn't do hacking whilst on the clock
[16:01] <Laney> oops
[16:03] <cjwatson> bah, libnet-server-perl failed to build
[16:09] <tkamppeter> when is beta2 freeze?
[16:09] <micahg> tkamppeter: tomorrow
[16:09] <directhex> Laney, can you take care while i go sit in a corner & sob?
[16:10] <tkamppeter> is missing on https://wiki.ubuntu.com/OneiricReleaseSchedule
[16:10] <micahg> skaet: ^^
[16:19] <skaet> micahg, tkamppeter - thanks.  Fixed.
[16:20] <Laney> directhex: don't think i will have time today
[16:20] <directhex> blarg
[16:20] <Laney> should just be one line though
[16:20] <directhex> at least we know the build passes :/
[16:20] <Laney> export MONO_DEBUG=1
[16:20] <Laney> put it in for the next normal upload if you don't want to bug lamont again
[16:21] <tkamppeter> skaet, thanks, another bug: beta1 is alink, beta2 not.
[16:21] <directhex> Laney, i guess put it in the .6 upload?
[16:22] <charlie-tca> tkamppeter: they don't become links until after they release
[16:22] <skaet> tkamppeter, actually that link tends to get added after the release is out.  Its a pointer to the release notes.  :)
[16:24] <Laney> directhex: yeah, if meebey does it soon. we can ask for it to be done on vernadsky again.
[16:24] <Laney> b2 freeze is tomorrow though ...
[16:27] <lamont> directhex: does that mean it failed in a way you didn't want?
[16:27] <Laney> forgot to export MONO_DEBUG to get the debugging info upstream wanted
[16:27] <directhex> lamont, it means it succeeded whereas previously it would have failed... but i forgot to enable the debug spew, because i'm an idiot
[16:28] <lamont> lol
[16:28] <lamont> if you want me to toss it through again, I'm fine with that.
[16:34] <ogra_> funny, must be the first time i see someone asking for an FTBFS on purpose :)
[16:34] <slangasek> wendar: is there something yet to be done wrt DEX for https://blueprints.launchpad.net/ubuntu/+spec/foundations-o-python-versions ?
[16:34] <Laney> imagine a build failure that only happens on two boxes, and imagine that both of those boxes are ubuntu buildds
[16:34] <Laney> yeah...
[16:35] <SpamapS> slangasek: so I've been monkeying with mapping's..
[16:35] <SpamapS> slangasek: seems like ifquery doesn't work with them as I'd expect it to..
[16:36] <slangasek> SpamapS: confirmed; I got halfway down that rabbit hole in the code yesterday before I gave up
[16:36] <slangasek> if only ifupdown was written in C, I'd probably have it done already
[16:36] <SpamapS> slangasek: Ok, so in light of that, we can take two stances. 1) we can add 'iface' to our parser that we already have in shell for 'auto' ...
[16:37] <SpamapS> slangasek: amen brother.. webm terrifies me
[16:37] <slangasek> 1) no no
[16:37] <SpamapS> slangasek: 2) we can just say that mappings will be waited upon for 2 minutes
[16:37] <slangasek> no, as discussed yesterday we should fix ifquery to actually dtrt with mappings
[16:37] <SpamapS> honestly, what is the server use case for mappings ?
[16:38] <slangasek> but you don't need to block on it, certainly
[16:38] <micahg> cjwatson: is this anything I need to worry about/file a bug for? http://paste.ubuntu.com/689338/
[16:38] <slangasek> SpamapS: so if you want to go for 2) with ifquery, that's perfectly fine with me
[16:38] <slangasek> and then the ifquery bug can be on my shoulders... or on smoser if he's keen to take it :)
[16:39] <SpamapS> slangasek: did you open a bug yet, otherwise I'll open one.
[16:39] <smoser> SpamapS, so what is wrong with ifquery ?
[16:39] <smoser> because i think it is right in the example you gave yesterday.
[16:39] <wendar> slangasek: it's been adopted as a DEX project, so the workitem is DONE
[16:40] <slangasek> smoser: ifquery doesn't do anything at all with mappings
[16:40] <smoser> i dont think it matters.
[16:40] <SpamapS> smoser: it is, but that example wasn't useful. ;)
[16:40] <wendar> slangasek: there's still work to be done on the bugs, but our goal was to contribute, not close them all down
[16:40] <slangasek> wendar: cool, ticking the box :)
[16:40] <SpamapS> http://paste.ubuntu.com/689342/
[16:41] <slangasek> smoser: right, for the server case the mapping bug probably doesn't matter.... I'm going to try to get it fixed in ifquery all the same
[16:41] <SpamapS> smoser: this interfaces file (/map just echos 'FOO') doesn't show eth0
[16:41] <SpamapS> So actually.. we can use ifquery, but note that mappings will *not* be waited on
[16:42] <wendar> hmm... somethings up with bugs.debian.org
[16:42]  * SpamapS throws ETOOMUCHINDIRECTION
[16:43] <smoser> SpamapS, right, and eth0 would be auto'd on that.
[16:43] <smoser> SpamapS, so do you want me to fix the upstart script to use ifquery ?
[16:45] <slangasek> I do :-)
[16:46] <SpamapS> smoser: to be clear, eth0 with that interfaces file is brought up at boot time, but is not shown in ifquery -a --list .. so it would not be waited on.
[16:46] <SpamapS> I'm *OK* with that
[16:46] <SpamapS> because I don't see a huge use case in mappings
[16:47] <SpamapS> its like a poor man's NetworkManager
[16:47] <smoser> right.
[16:47] <smoser> ifquery shoudl be fixed, and we should open a bug
[16:47] <smoser> but yeah.
[16:48] <SpamapS> ok, I'll open said bug
[16:48] <SpamapS> Yay throw it on the pile.. oi
[16:48] <slangasek> throw it on my pile
[16:48] <slangasek> (i.e., assign)
[16:48] <SpamapS> ack
[16:48] <smoser> i'll fix ifup upstart job to use 'ifquery --list'
[16:48] <smoser> slangasek, SpamapS is that the "right" arguments ?
[16:49] <smoser> as ifquery sparse documentation, hard to decide what is right
[16:49] <smoser> seems to me that is right.
[16:50] <SpamapS> I think it defaults to --allow auto, but I would explicitly pass that
[16:50] <cjwatson> micahg: dunno, where's grub-pc configured to install GRUB to?
[16:51] <slangasek> SpamapS: yes, please explicitly pass that, because I'm planning on fixing that in ifquery :)
[16:51] <micahg>  cjwatson: no idea, I haven't modified this AFAIK
[16:51] <slangasek> smoser: ^^
[16:51] <cjwatson> micahg: debconf-show grub-pc
[16:52] <directhex> lamont, i added the env var to https://launchpad.net/~directhex/+archive/monoxide?field.series_filter=oneiric
[16:52] <micahg> cjwatson: it has a /dev/disk/by-id output
[16:52] <smoser> slangasek, ok.
[16:54] <chrome_> !release
[16:55] <SpamapS> slangasek: bug 850166 opened for you
[16:56] <cjwatson> micahg: yes.  what exactly does it say though?
[16:57] <micahg> cjwatson: /dev/disk/by-id/ata-ST9500420AS_5VJBCYN8
[16:58] <cjwatson> micahg: ok, so it's managed to work around the problem but only in a fragile way because your partitioning is bad
[16:58] <cjwatson> micahg: there should be at least 63 sectors before the first partition (nowadays, 1MiB is usuall)
[16:58] <cjwatson> *usual
[16:58] <cjwatson> micahg: it may be hard for you to change this now, of course
[16:58] <micahg> yeah, the drive came preinstalled
[16:58] <micahg> cjwatson: so, you don't need a bug report then I take it?
[16:59] <cjwatson> micahg: nope
[16:59] <micahg> cjwatson: thanks :)
[16:59] <lamont> directhex: building on vernadsky: wtf2
[16:59] <lamont> afk for a few
[17:00] <directhex> thanks :)
[17:01] <smoser> SpamapS, i think that bug is moderately bogus
[17:02] <smoser> as 'ifquery' is then going to have to run code (/map) that is only expected to be run at ifup time
[17:03] <smoser> i guess its not so bad. but kind of wierd
[17:04] <SpamapS> Its a little weird, and might scare some people, but the scripts are meant to map every time ifup is run. Honestly, I'd be happier with just *ditching* mappings and saying thats what connman or network-manager is for.
[17:05] <SpamapS> But easier, right now, is for us to just say that they won't be waited on.
[17:07] <apw> pitti, do you have any binary.debs for the plymouth flicker issue
[17:08] <apw> pitti, am trying to do some testing here
[17:09] <chrisccoulson> jono - do you have evolution-couchdb installed btw?
[17:10] <jono> chrisccoulson, aha!
[17:10] <jono> installing now
[17:10] <chrisccoulson> heh :)
[17:10] <chrisccoulson> jono - it stil doesn't work though. i get a new error from the eds couchdb backend now :(
[17:10] <jono> I think I uninstalled it as rodrigo suggested to get rid of the constant errors
[17:10] <jono> well test now
[17:10] <jono> chrisccoulson, ahh
[17:11] <jono> thanks chris
[17:11] <jono> thanks chrisccoulson
[17:11] <chrisccoulson> np :)
[17:12] <jono> chrisccoulson, I am getting an error about not opening ubuntu one address bug because of the process
[17:13] <chrisccoulson> jono - i get "Cannot open book: Could not create DesktopcouchSession object"
[17:13] <chrisccoulson> do you get something similar?
[17:13] <jono> chrisccoulson, nope
[17:13] <chrisccoulson> that comes from the evolution-couchdb backend, and suggests that desktopcouch is still broken for mw
[17:13] <chrisccoulson> **me
[17:13] <jono> There was a problem opening the address book "Ubuntu One" - the message returned was: Cannot open book: Cannot process, book backend is opening
[17:13] <jono> that is what TB tells me
[17:14] <chrisccoulson> jono - perhaps try killing e-addressbook-factory, restarting thunderbird and trying again. maybe it needs a restart to load the new backend
[17:16] <jono> chrisccoulson, killed it, now I don't see an error but no contacts in the address book
[17:20] <chrisccoulson> jono - yeah, it's broken here too. but it's also broken in evolution :/
[17:22] <jono> chrisccoulson, good luck, pal
[17:22] <chrisccoulson> heh :)
[17:22] <chrisccoulson> thanks
[17:22] <jono> it would suck if we have another release with broken U1 contacts
[17:22] <chrisccoulson> was it broken last cycle too?
[17:23] <chrisccoulson> i have to admit, i've never used it before because couchdb was the first thing i uninstalled after doing a fresh install ;)
[17:25] <jono> chrisccoulson, it kind of worked
[17:25] <jono> for a large number of contacts it broke
[17:29] <dobey> hmm
[17:30] <dobey> jono: what version of couchdb-bin do you have installed?
[17:31] <dobey> oh
[17:31] <dobey> nevermind
[17:31] <lamont> directhex: lol
[17:32] <lamont> directhex: let me know when you're ready for wft3
[17:32] <directhex> dare i ask?
[17:32] <lamont> make[9]: Entering directory `/build/buildd/mono-2.10.5/mcs'
[17:32] <lamont> Invalid option for the MONO_DEBUG env variable: true
[17:32] <lamont> Available options: 'handle-sigint', 'keep-delegates', 'reverse-pinvoke-exceptions', 'collect-pagefault-stats', 'break-on-unverified', 'no-gdb-backtrace', 'dont-free-domains', 'suspend-on-sigsegv', 'suspend-on-unhandled', 'dyn-runtime-invoke', 'gdb', 'explicit-null-checks', 'init-stacks'
[17:33] <directhex> ._.
[17:33] <directhex> bloody computers
[17:36] <jono> dobey, :-)
[17:37] <dobey> hrmm, evolution is weird
[17:38] <SpamapS> anybody know how to get skype working on oneiric?
[17:38] <SpamapS> do I have to download the skype from their site now? :(
[17:39] <micahg> SpamapS: add natty partner repo, install skype:i386?
[17:41] <SpamapS> no
[17:41] <SpamapS> I forgot to add the foreign arch
[17:41] <SpamapS> had already tried that.. :p
[17:43] <dobey> hrmm
[17:43] <directhex> lamont, wtf3 just uploaded to my ppa.
[17:43] <dobey> skype + unity do not mix
[17:44] <micahg> SpamapS: wfm
[17:44] <SpamapS> micahg: its not necessary tho
[17:45] <micahg> SpamapS: what's not necessary?
[17:45] <SpamapS> micahg: its in oneiric's partner repo.. just need the i386 foreign arch
[17:45] <micahg> SpamapS: no, it's not in oneiric's partner yet
[17:45] <dobey> SpamapS: not according to my apt-cache :)
[17:46] <SpamapS> really? hrm.. this is the suck
[17:47] <ScottK> partner is rarely populated much before release.
[17:47]  * SpamapS should just abandon skype anyway.. :-P
[17:48] <dobey> SpamapS: is there a version of skype that works right with unity?
[17:48] <SpamapS> It worked fine w/ unity 2d .. haven't been able to run it in a couple of weeks.. and since then I switched back to regular unity
[17:50] <dobey> SpamapS: is there some indicator that handles the old notification area icons that's not installed by default in oneiric?
[17:51] <SpamapS> err, they were whitelisted in natty
[17:51] <micahg> AFAIK, the notification whitelist was dropped for oneiric
[17:51] <dobey> there is definitely no icon on my panel in oneiric
[17:52] <dobey> i closed the skype window and got left with a process with no UI :)
[17:52] <dobey> of course, also a bug in skype, but i can't go patch it :P
[17:52] <SpamapS> awesome, so enabling i386 as a foreign arch has made my i386-only printer drivers cause huge issues
[17:52] <SpamapS> now apt wants to remove libc-bin
[17:55] <ScottK> dobey: It's not really a bug in Skype.  It's Ubuntu being different than every other Linux distro and expecting the entire FOSS world to follow.
[17:55] <ScottK> KDE implemented the same notification area technology, but with a fallback.
[17:56] <ScottK> It's a deliberate choice to have issues like this.
[17:57] <dobey> ScottK: no. it's a bug in skype if it doesn't properly handle the notification area not being there. if we have something running that is embedding the icon and just not showing it on-screen though, i would be inclined to agree
[17:58] <ScottK> Whatever.
[17:58] <barry> cjwatson, slangasek: so, by disabling python-docutils in python-defaults, i'm making progress again.  too early to declare victory, but i'm past the latest roadblock
[17:59] <micahg> barry: was there a reason why python-debian didn't get ported to dh_python2 for beta1?
[17:59] <barry> micahg: no reason i can think of except that perhaps it didn't make it onto the list of cd packages?
[17:59] <barry> micahg: does it still need porting?
[18:00] <micahg> barry: yes, Debian did it in the latest upload, but there are quite a few changes, I didn't know if they were safe at this point since that library is used heavily
[18:00] <barry> micahg: let me take a quick look
[18:01] <micahg> barry: I have a task open for the xubuntu CD conversion bug, feel free to take it
[18:01] <barry> micahg: bug #?
[18:01] <micahg> bug 847514
[18:01] <barry> micahg: got it
[18:02] <micahg> barry: thanks!
[18:02] <dobey> ScottK: it has the same problem under gnome-shell. it doesn't properly handle the notification area icon not being embedded. -> bug in skype. :)
[18:04] <ScottK> When infrastructure changes under an app with no fallback, I don't think that's a bug in the app.  It's poor project planning on the infrastructure's part, but there's not much chance of either of us fixing it, so we can agree to disagree.
[18:09] <smoser> SpamapS, slangasek thinking more about that bug in ifquery...
[18:10] <smoser> the way we're using it, not only would ifquery call the mappings scripts on 'list' (when normally they'd only be expecting to be called on ifup), but also we're going to be basically calling them recursively.
[18:10] <[swe]jeppe> hi everyone
[18:10] <smoser>  - ifup -a
[18:11] <smoser>   - mapping script called
[18:11] <SpamapS> and returned from
[18:11] <smoser>   - upstart job calls ifquery -a --list
[18:11] <smoser>   - ifquery calls mapping script
[18:11] <SpamapS> repeatedly, but not recursively
[18:11] <smoser> yeah
[18:12] <smoser> SpamapS, bug 850226 opened.
[18:13] <SpamapS> smoser: oh good thanks. :) I meant to do that yesterday
[18:13] <[swe]jeppe> can someone tell me whats the biggest obstickle why its not working to play games on linux?
[18:14] <dobey> [swe]jeppe: i don't think this is the channel for answering that, but i presume the answer is "you haven't installed them" :)
[18:14] <Pici> [swe]jeppe: #ubuntu-offtopic is the channel for general discussion.
[18:16] <[swe]jeppe> no, im just woundering whats block when u install games on linux. hard to explain in english
[18:16] <barry> micahg: how truly odd.  bug 788514 has a task for python-debian, which is marked fixed released, and i have a branch sitting here with the necessary changes.  i usually never manually flip the fix released status, so i'm not sure what's going on.  maybe i never uploaded the change.  anyway, i'll do that now
[18:17] <Pici> [swe]jeppe: And this channel isn't for general disucssion, its a working development channel.
[18:18] <[swe]jeppe> ok sorry will move :-)
[18:21] <hyperair> hi. any ubuntu-release members around?
[18:21] <smoser> slangasek, around ?
[18:21] <smoser> so if i pass '--allow all' to 'ifquery --list', then it is busted at the moment.
[18:21] <jdstrand> @pilot in
[18:22] <smoser> never mind.
[18:22] <smoser> dugh
[18:23] <slangasek> smoser: --allow auto :)
[18:23] <smoser> it is also broken if i pass --allow moser-youre-an-idiot
[18:23] <smoser> strange
[18:24] <slangasek> smoser: as for ifquery calling the mapping scripts... the design says it's fine to call the mapping script as long as the interface is down, but not while it's up, but we don't need to call it while up to get the mapping because that's in the ifstate file... this is the bit that needs stitching together, and where I threw my hands up in irritation with noweb last night
[18:24] <smoser> sweet sweet noweb
[18:25] <slangasek> SpamapS: what i386-only printer drivers are giving you problems?
[18:28] <SpamapS> slangasek: 3rd party
[18:28] <SpamapS> slangasek: I had to force install them long long ago
[18:29] <SpamapS> slangasek: canon pixma mp560 to be specific
[18:30] <skiy1337> I've just written a piece of software. If you have an FTP server, it let's you store you documents on it.
[18:30] <SpamapS> slangasek: it seems to be something else actually.. I removed them and apt still wants to uninstall libc-bin when I try to install skype:i386
[18:31] <slangasek> SpamapS: oh, lovely
[18:31] <hallyn> lp:ubuntu/libvirt is just an alias to lp:ubuntu/oneiric/libvirt right?
[18:31] <slangasek> SpamapS: any better with 'apt-get install skype:i386 libc-bin'?
[18:32] <hallyn> (quick test says 'yes' :)
[18:32] <SpamapS> slangasek: yes that works.. but.. wtf?
[18:32]  * SpamapS wonders why gcc-4.6-base:i386 is needed for skype.. :-P
[18:33] <slangasek> SpamapS: because gcc-4.6-base is needed by libgcc1 is needed by $world :)
[18:33] <SpamapS> lurvely
[18:33] <slangasek> gcc-4.6-base is really very base
[18:33] <slangasek> what's the delta of installs/removals look like between the "I want to remove libc-bin" and "ok boss" cases?
[18:33] <SpamapS> wow.. one gets a week behind on upgrades and oneiric has replaced nearly everything
[18:34] <SpamapS> slangasek: very similar, just this time it doesn't want to remove libc-bin
[18:35] <slangasek> heh
[18:36] <slangasek> then it's probably picking a bad upgrade path on account of eglibc being out of date; I've seen that a couple of times before
[18:36] <slangasek> haven't gotten around to filing a bug for it yet, maybe you could?
[18:41] <SpamapS> slangasek: a bug in apt?
[18:41] <slangasek> SpamapS: yep
[18:42]  * SpamapS looks through log to find what version of libc-bin was replaced
[18:43] <SpamapS> Preparing to replace libc-bin 2.13-17ubuntu2 (using .../libc-bin_2.13-20ubuntu2_amd64.deb) ...
[18:44] <SpamapS> slangasek: for the record, aptitude also just refused to do anything
[18:44] <slangasek> how sensible of it :)
[18:48] <SpamapS> alright, filed bug 850264 with, hopefully, enough of the details
[18:50] <slangasek> SpamapS: looks good, thanks
[19:14] <micahg> barry: ok, thanks much
[19:14] <barry> micahg: no worries
[19:46] <lamont> (took 1 hour, 34 minutes, 49.9 seconds) <-- micahg
[20:26] <Amoz> notice like a boss
[21:04] <hallyn> regcomp succeeds on "^\\s*(\\S+),(\\S*),(\\S+),(\\S+)\\((\\S+)\\),(\\S+),([0-9]+),?\\s*$" , but fails on "^(?:\\s*(\\S+))?\\s*(\\S+),(\\S*),(\\S+),(\\S+)\\((\\S+)\\),(\\S+),([0-9]+),?\\s*$" (identical except for the &(?:\\s*(\\S+))? at front).
[21:05] <hallyn> I assume it's failing on the '?:'.  Can anyone confirm if that is not supported, and if there's another way to do that?  (namely, match the \s* if it is there, but not return it as a match)?
[21:06] <hallyn> maybe i should just do '.*'...
[21:23] <slangasek> hallyn: isn't that specific to perl regexp, not posix regexp?
[21:27] <hallyn> slangasek: sadly it would seem so.  So is there any posix way to do that?
[21:27] <hallyn> ATM I'm resorting to hacking the c code around it to discard the prefix if present
[21:27] <hallyn> but ideally I could do it in the regexp
[21:31] <broder> hallyn: you could look at the number of groups you get back, if that's the only one followed by a ?
[21:32] <broder> (sorry, there's a bit too much punctuation in there for me to actually parse what's going on)
[21:32] <hallyn> broder: the structure of the existing code is such that one function passes the regex and # expected fields into another  :(
[21:32] <hallyn> now if i could have it return an empty match i suppose that would be fine
[21:33] <broder> yeah, i can't come up with any way to do that
[21:33] <hallyn> actually,
[21:34] <broder> well...i guess you might be able to do something like (|\\s*(\\S+))
[21:34] <broder> with no ?
[21:34] <hallyn> that code does take multiple regexes, so maybe i could just pass 2
[21:34] <hallyn> yeah that might work too ('|')
[21:34] <broder> you'd want to be sure to drop the group around the inner \\S+
[21:34] <hallyn> thanks - lemme see if what i'm doing now works, and then maybe i'll try that
[21:35] <broder> it's not necessary, and changes the number of groups based on the matched expr
[21:35] <hallyn> you mean '(|\\s*\\S+)' then?
[21:36] <broder> yeah
[21:36] <hallyn> ok thanks
[21:36] <hallyn> (gotta run, may try that later)
[21:48] <dupondje> how is it possible to have this: LANGUAGE=nl:zh_CN:en_AU:en ?!
[21:48] <dupondje> selected nl in the control panel, and made it a weird combo ...
[21:55] <dobey> jdstrand: hi. care to prioritize sponsoring of https://code.launchpad.net/~dobey/ubuntu/oneiric/ubuntuone-installer/release-174/+merge/75437 so it gets approved before string freeze? :)
[22:04] <jdstrand> I'll look at it
[22:08] <jdstrand> dobey: this needs a new orig.tar.gz
[22:10] <jdstrand> dobey: can you provide me with a signed source package?
[22:12] <dobey> jdstrand: huh, i've not had anyone say this before for merges into lp:ubuntu/ branches. where should i put the source package?
[22:13] <jdstrand> dobey: well, it is a new upstream release and I want to be able to verify the checksums on the orig.tar.gz with the upstream source
[22:13] <jdstrand> dobey: you drew the short straw getting a security/archiveadmin member :P
[22:14] <jdstrand> dobey: chinstrap is fine
[22:15] <dobey> jdstrand: https://chinstrap.canonical.com/~dobey/
[22:15] <YokoZar> slangasek: uploading fixed ssl ia32-libs now...probably finish uploading in a few hours :D
[22:15] <jdstrand> dobey: thanks, I'll process it now
[22:16] <dobey> jdstrand: thanks
[22:17] <slangasek> YokoZar: did it require anything more than the refresh?
[22:17] <YokoZar> slangasek: I had to add SSL 1.0.0 in there as well
[22:18] <YokoZar> (for Wine)
[22:18] <jdstrand> dobey: can you adjust the perms of the dsc and debian.tar.gz files?
[22:18] <slangasek> YokoZar: hmm.  how/why does it depend on the newer ssl?  I'm rather unhappy with the idea of having two copies of libssl in ia32-libs - one is quite bad enough
[22:19] <YokoZar> slangasek: because ssl-dev requires the 1.0.0 copy and Wine is building against that
[22:19] <dobey> jdstrand: fixed, sorry
[22:19] <jdstrand> thanks
[22:21] <slangasek> YokoZar: does it need headers or just the .so?  Far better to manually add the .so symlink if that's all it needs
[22:25] <YokoZar> slangasek: I'm not sure it's good to open the can of worms of building with newer header files than the actual linked library...
[22:25] <slangasek> certainly not, that's why I asked if it needs headers
[22:26] <YokoZar> Yeah it does
[22:26] <slangasek> sigh. ok.
[22:36] <jdstrand> @pilot out
[22:51] <bdrung> jamespage: around?
[23:40] <doko> "Gentoo sanity check failed!" does this even exist?
[23:42] <desrt> doko: ;)
[23:42] <broder> duh, why do you think it faileD?
[23:43] <doko> bug 770749