[02:06] <mwhudson> argh
[02:07] <mwhudson> is there some way i can easily hack about in a revision controlled dir (bzr or git i don't mind) and turn my local commits into patches in debian/patches?
[02:07] <mwhudson> (only the parts of my local commits that don't touch debian/ obv.)
[02:09] <sarnold> mwhudson: it's not what you asked for, but 'quilt shell' may come close to giving you something that works
[02:09] <mwhudson> erm
[02:09] <mwhudson> i guess so
[02:43] <darkxst> mwhudson, you can use gbp-pq to create a branch with all debian patches applied as commits
[04:31] <hallyn> sarnold: are you by chance an apport guru?
[05:30] <jakesyl> hey guys i've been looking to help out with ubuntu dev and already know a considerable amount of python, but what other languages should i learn? x86?
[05:31] <RAOF> jakesyl: If by “x86” you mean “x86 assembler”, then that's a pretty low-return investment of learning :)
[05:32] <RAOF> jakesyl: The answer, as with everything open-source is: what do you want to do? ☺
[05:32] <jakesyl> i learn for the sake of learning, and of course applying my learning to a worthy cause, so what can i learn to be of most assistance RAOF
[05:33] <RAOF> Because that will determine what you'll need to know. Want to hack on GNOME Do? You'll need to know (or learn) C#. Lots of the Ubuntu tools are written in Python, so you could help there already; Ubuntu Touch uses QML / Qt, for which a dialect of C++ can be useful.
[05:34] <RAOF> It's sort of like going up to a science department and saying “I'd like to help in science; what can I learn to be most useful” :)
[05:34] <StevenK> RAOF: Install a power-mad AI and get her to do the science, obviously.
[05:35] <RAOF> Generally the best way to help is to find something that doesn't work the way you want, or is broken in some way, and fix it :)
[05:36] <jakesyl> eh RAOF, hypothetically if i wanted to build an ubuntu like os from scratch, what would i need to know
[05:36] <rww> StevenK: I vaguely recall popular culture telling me that this ends badly with cake and lies and such.
[05:37] <StevenK> rww: :-)
[05:37] <jakesyl> big hypothetical
[05:37] <RAOF> jakesyl: Depends on what you mean by both “from scratch” and “Ubuntu-like OS”.
[05:38] <jakesyl> RAOF, i mean completely from scratch, as in an ubuntu iso
[05:38] <jakesyl> and general linux
[05:38] <RAOF> Well, we've got tools to do that :)
[05:38] <RAOF> But from the raw upstream source you'd need to know a whole bunch of system administration, because you're going to be running a build farm.
[05:38] <jakesyl> what languages?
[05:39] <RAOF> That's kindof the wrong question :)
[05:40] <jakesyl> hmm why?
[05:40] <RAOF> But you'll be dealing with everything from raw assembly of whichever architectures you're interested in to compilers for the same, to C, C++, python, ruby, perl, C#, Java...
[05:41] <jakesyl> RAOF, perhaps a little context, i'm 15 i've been writing perl since i was 8 finally decided it was a dead language, i still wrote python wrappers in perl, and i'm just now realizing i should move on
[05:43] <RAOF> jakesyl: The most useful way to learn is to find a (coding) project that needs help (which is basically everything!) and work on it. Which project you help determines what you need to learn.
[05:43] <RAOF> (Although the chances of needing to learn x86 assembler are very low ☺)
[05:44] <jakesyl> so what would i need x86 for
[05:45] <RAOF> Work on compilers or runtimes, mainly.
[05:47] <jakesyl> RAOF, the reason i ask is ever mentor i've ever had brags about memorizing hex tables
[05:48] <jakesyl> hmm RAOF, if i wanted to do a ui/ux overhaul and just make it look cleaner, what would i need to learn?
[05:49] <RAOF> Those sound like some sucky mentors :(
[05:49] <RAOF> So, choices!
[05:49] <jakesyl> well either that or a universal package manager
[05:49] <RAOF> If you want to work on the actual UI toolkits - that'd be Qt and GTK - you'll be wanting C++ and C(ish)/Vala, respectively.
[05:50] <RAOF> To work on applications, you get to pick an application and find out what it uses :)
[05:50] <jakesyl> but there's no uni package manager is there?
[05:50] <rww> there are too many universal package managers. someone is going to end up making a universal universal package manager
[05:51] <jakesyl> rww, i've never seen a seamless one tho, now i'm not an apple fanboy, i'm on ubuntu, but the seamless integration is what i want, click on the package it opens, is there anything like that?
[05:51] <RAOF> Yes; gdebi
[05:52] <RAOF> Or do you not want to install things?
[05:52] <jakesyl> why doesn't it come pre installed?
[05:52] <RAOF> It does.
[05:52] <RAOF> (I think?)
[05:52] <jakesyl> so how come when i click an application it doesnt just open
[05:52] <rww> RAOF: it's in universe, so I assume not
[05:53] <RAOF> Oh, of course. Software Centre does that now.
[05:53] <RAOF> jakesyl: What do you mean by ‘an application’? :)
[05:53] <Mirv> if a core dev is around, unity8 CI Train release would need a packaging ack for their debian/ changes https://ci-train.ubuntu.com/job/landing-005-2-publish/lastSuccessfulBuild/artifact/packaging_changes_unity8_8.00+14.10.20140814.1-0ubuntu1.diff
[05:53] <RAOF> We *don't* have application bundles ala OS X, but that's not a technical problem.
[05:53] <geofft> I think the click-packages stuff is also relevant here?
[05:54]  * RAOF looks at that packaging diff
[05:54] <jakesyl> metasploit is a great example, i download metasploit i hit the package it should auto install and i understand how to do it but it should be seamless should it not?
[05:55] <RAOF> If you download an Ubuntu package and double click on it it *will* install (or, rather, it'll come up in the Software Centre asking if you want to install it)
[05:55] <jakesyl> okay so i guess i'll focus on ux
[05:59] <RAOF> Mirv: That looks fine.
[05:59] <Mirv> RAOF: thanks!
[06:07] <sarnold> hallyn: sorry, I don't know much about apport
[06:22] <dholbach> good morning
[06:26] <dholbach> does anyone of the Canonical folks know if there's an arm64 porter box?
[06:32] <slangasek> dholbach: there is not; there's an open RT for one
[06:32] <dholbach> all right, thanks
[06:32] <slangasek> (there are physical boxes but they're not yet provisioned as a porter box)
[07:12] <LocutusOfBorg1> hi dholbach, were you looking for the arm64 porter because of wxwidgets3.0?
[07:12] <LocutusOfBorg1> (the build has been successful)
[07:12] <LocutusOfBorg1> thanks ;)
[07:12] <dholbach> LocutusOfBorg1, indeed I was - in the end I trusted your build log and knew I could find you if things exploded :)
[07:12] <LocutusOfBorg1> oh yes ;)
[07:13] <LocutusOfBorg1> feel free to mail me, or to find me there, and the fix was quite trivial anyway (disable the precompiled headers), fortunately the bug has been fixed in gcc
[07:14] <dholbach> cool
[07:15] <LocutusOfBorg1> BTW I never been able to reproduce in pbuilder/arm64 so I guess the old buildd lp system was just buggy, again a great thanks to cjwatson :D
[07:23] <tjaalton> is there a way to force dch to not append the ubuntu version?
[07:23] <tjaalton> in a conffile
[07:26] <LocutusOfBorg1> tjaalton, the man says probably DEBCHANGE_VENDOR
[07:29] <LocutusOfBorg1> "dch --vendor debian"
[07:30] <tjaalton> hm indeed
[07:30] <tjaalton> just need an alias then
[07:31] <LocutusOfBorg1> I think so and I think I'll need it too ;)
[07:31] <LocutusOfBorg1> it is annoying sometimes :)
[07:31] <tjaalton> especially when you later notice that debcommit pushed it to $vcs.. to a debian branch
[07:35] <LocutusOfBorg1> argh
[07:50] <LocutusOfBorg1> dholbach, I opened two bugs for you ;) sorry
[07:53] <dholbach> LocutusOfBorg1, not necessarily just for me, right? :)
[08:07] <LocutusOfBorg1> no dholbach :) but you rock in merging/syncing lol
[08:08] <dholbach> haha
[08:08] <dholbach> I need to take care of a few other things first - if nobody beats me to it, I might have a look later on
[08:08] <LocutusOfBorg1> wonderful!!!
[08:08] <LocutusOfBorg1> I really would like to fix the wx2.8 and wxpython3.0 stuff
[09:35] <doko> dobey, ubuntu-sso-client finally \o/
[09:40] <Laney> doko: any objection to taking pcre3 8.35?
[09:40] <doko> Laney, not at all
[09:42] <Laney> cool
[09:42] <Laney> infinity: your merge, can I steal it?
[09:44] <infinity> Laney: My merge of what?
[09:45] <infinity> Laney: Oh, pcre3?  Go nuts.
[09:45] <Laney> Tah
[10:10] <rbasak> xnox: bug 1355992 claims the last samba merge inadvertently dropped an upstream cherry-pick. Please could you take a look?
[10:32] <halfie> sudo aa-genprof /usr/bin/vim results in "TypeError: 'bool' object does not support item assignment"
[10:32] <halfie> the AppArmor user space seems to be in a terrible state
[10:33] <rbasak> I think the genprof issue is known about
[10:33] <rbasak> AIUI, it's limited to genprof.
[10:34] <rbasak> jjohansen: ^^
[10:35] <halfie> I thought that AppArmor was like a bullet-proof production ready thing ;(
[10:35] <halfie> I guess I will have to go with SELinux (on Ubuntu!)
[11:10] <lullis> Hi, all. I have a few questions regarding (I think) lightdm customization. Basically, I am working on a system that needs to provide remote authentication. And to do that, the user needs to be able to (1) set up the network connection and (2) be able to also "register" directly on the greeter screen.
[11:11] <lullis> My questions would be: regarding (1), can I add the NM applet to top bar of any the (gtk/unity) greeters?
[11:19] <lullis> regarding (1) and (2), I have a python/GTK application that handles the setting up a new account on the remote server. How hard would it be to change the C greeters to add a handler that starts my python code?
[11:19] <doko> jpds, strongswan ftbfs on armhf and powerpc. any reason not to use libgcrypt20?
[11:20] <lullis> And also, is any better place to ask these questions?
[11:22] <rbasak> halfie: AppArmor is production ready, and has been used in production for many years. I wasn't aware you were talking of a production release. No idea about that, but genprof is a helper and not a production piece.
[11:22] <rbasak> halfie: try #ubuntu-hardened for apparmor discussion. The security folks listen in in there.
[11:48] <jpds> doko: Upstream still has to figure it out: https://wiki.strongswan.org/issues/674
[12:07] <doko> jpds, please can you try building with -O2 on ppc64el?
[12:07] <jpds> doko: Ah, upstream just pinged me and we were doing a git bisect on a porter box.
[12:08] <doko> jpds, but up to now, all test failures seen with -O3 did point to invalid source code ...
[12:10] <jpds> doko: Upstream says -O2 is the default in the configure script.
[12:11] <doko> jpds, and?
[12:12] <doko> ppc64el is built with -O3 as the default
[14:55] <halfie> hi, I am trying to download all utopic package and the debug information packages for some offline analysis. how do I go about it? Currently I am using debmirror.
[15:00] <rbasak> halfie: debug packages are at ddebs.ubuntu.com. The rest is as your sources.list says for utopic.
[15:01] <halfie> rbasak, ah, I see. is there a way to mirror the debug packages for utopic only? I guess I could try using debmirror for it?
[15:02] <rbasak> halfie: just mirror ddebs.ubuntu.com I guess?
[15:02] <halfie> rbasak, it will be too large :)
[15:03] <rbasak> halfie: well, the rest is up to your mirroring tool. debmirror can filter as needed.
[15:03] <halfie> ok, giving it a shot now
[15:06] <tedg> ev, Would you be opposed to us not filtering fields on recoverable errors in Whoopsie?
[15:13] <lamont`> has anyone run into grub_isprint not found (and the subsequent bail into grub rescue) on a machine upgraded from saucy to trusty? (did this last night, not so wonderfull)
[15:13] <lamont`> didn't find anything in launchpad on the subject
[15:24] <lamont`> nm.
[15:26] <ev> tedg: very much opposed, given that people shove entire log files into these fields and cassandra does not do blob storage
[15:26] <ev> what's the motivation?
[15:27] <tedg> ev, Well for instance I have UAL have a recoverable error for "bad AppID" and then it has a custom field with the appid that was bad. Not getting that custom field.
[15:27] <tedg> ev, https://errors.ubuntu.com/bucket/?id=ubuntu-app-launch-invalid-appid
[15:28] <tedg> ev, Can we switch to Mongo? I hear it has all the features.
[15:29] <tedg> ev, Seems like we could filter by size instead of a white list?
[15:35] <ev> suggest mongo again and I'll come find you
[15:35] <ev> tedg: sometimes we have big fields we want to carefully split
[15:35] <ev> like Stacktrace :)
[15:36] <tedg> ev, Sure, so let the whitelist be to avoid the size limits not to be in/out.
[15:36] <tedg> ev, So, < 1KB or in whitelist
[15:36] <ev> so any field below a certain size plus these big ones we care about?
[15:36] <ev> yeah
[15:36] <ev> works for me
[15:36] <ev> I'll +1 that MP
[15:36] <ev> but remind me I said this :)
[15:36] <tedg> ev, K, I'll work on that MP
[15:37] <tedg> ev, I'll remind you daily :-)
[15:37] <ev> lol
[15:37] <tedg> ev, I don't know cassandra, is 1KB reasonable?
[15:39] <ev> yeah, I think the barrier (for Thrift) is 16 MB, but the other thing I'm trying to avoid is quickly growing the database when the IS team responsible for getting us out of the ENOSPC hole has been moved onto other top secret projects
[15:39] <ev> so we can start at 1K and increase from there
[15:40] <ev> tedg: http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Increasing-thrift-framed-transport-size-in-mb-td6825655.html
[15:41] <tedg> ev, Oooh, secret projects. 1KB I think works for my usecases.
[15:41] <ev> note that we're on cassandra 1.0.lulz, so anything you read about 1.2, 2.0 or whatever is a pipe dream at this point
[15:44]  * tedg wonders what else would break if an appid was 1KB in length
[16:54] <dpm> hi sarnold, would you mind doing a final review to the File Manager security checks MP? We've got the AP tests working now, and Arto did a couple of tweaks to the code. You gave your +1 already, but it'd be good to have a final sanity check after the last few changes
[16:54] <dpm> https://code.launchpad.net/~ubuntu-filemanager-dev/ubuntu-filemanager-app/require-screenlock-password/+merge/230058
[17:50] <tjaalton> shouldn't sbuild/schroot clean up stuff from /tmp too after exiting the chroot?
[18:13] <Unit193> cjwatson: Hello!  Not sure if you're back from VAC, but just a bump on tasksel refresh?
[18:46] <mdeslaur> doko: mind if I merge subversion?
[18:47] <doko> mdeslaur, not at all. can you apply the patch from 1353142 and revert to -O3?
[18:47] <mdeslaur> doko: sure
[18:51] <jtaylor> doko: mind if I do numpy?
[18:53] <doko> jtaylor, sure. did you see may scipy workaround?
[18:53] <jtaylor> no?
[18:53] <jtaylor> I wanted to check that out this weekend
[18:53] <jtaylor> also needs another fix for statsmodels
[18:54] <doko> are you at debconf?
[18:54] <jtaylor> no
[18:54] <doko> http://launchpadlibrarian.net/182175661/python-scipy_0.14.0-1_0.14.0-1ubuntu1.diff.gz
[18:55] <jtaylor> I wonder why that is failing, I did test i386 for the 0.14 upload
[18:57] <jtaylor> also seems to work in debian
[18:57] <jtaylor> maybe gcc-4.9?
[18:58] <jtaylor> I'll look at it later
[18:59] <doko> yep, but it's i386, and I didn't care too much, therefore the ignore
[19:00] <jtaylor> yes i386 fails numeric tests all the time
[19:04] <jtaylor> doko: I filed a bug about gcc dropping -ltsan https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62105
[19:04] <jtaylor> not much useful information yet but possibly we have to add --no-as-needed in the ubuntu package
[19:05] <doko> jtaylor, maybe, but since today we can use --push-state and --pop-state
[19:05] <doko> see the ld docs
[19:05] <jtaylor> uh thats neat
[19:29] <jtaylor> doko: in case you want to sponsor, bug 1357044
[20:37] <mwhudson> darkxst: hm, thanks
[20:38] <mwhudson> i found gbp completely confusing last time i looked at it, but gbp pq seems sensible on a very quick look...
[20:41] <doko> jtaylor, merged
[20:47] <jtaylor> thx
[22:45] <hallyn> so, i have /usr/share/apport/package-hooks/source_qemu.py , but it doesn't get called when qemu-system-x86 crashes (that is, i kill it manually with kill -11).  Am I supposed to do something else?  Register it somehow?