[06:59] <cpaelzer> good morning
[08:12] <Mirv> cjwatson: would https://code.launchpad.net/~timo-jyrinki/click/dont_use_@_in_testscontrol/+merge/302514 fix the autopkgtest failure?
[08:58] <jtaylor> LocutusOfBorg: I figured out why clang is not working in 16.04
[08:58] <jtaylor> having gccgo-6 installed breaks it
[08:58] <jtaylor> it selects gcc6 then but that installation is incomplete
[09:12] <xnox> mwhudson, doko, see jtaylor ^ sounds like fun
[09:13] <jtaylor> bug 1611676
[09:39] <doko> jtaylor, well, fix that silly clang ... there are absolutely no c++ headers
[09:47] <cjwatson> Mirv: approved, feel free to try it :)
[09:47] <Mirv> thanks!
[09:47] <cjwatson> seems plausible at least
[11:59] <doko> cjwatson, cyphermox: filed https://bugs.launchpad.net/ubuntu/+source/grub/+bug/1611740  I think I talked with cyphermox about it at the last sprint. do you can think of anything which still uses it?
[12:00] <showaz> ubuntu — clamav-unofficial-sigs v3.7.2 (updated 2013-08-25)
[12:00] <showaz> https://github.com/extremeshok/clamav-unofficial-sigs - 20 July 2016 (5.4.1)
[12:00] <Mirv> cjwatson: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#click :( now the configure is searching for packagekit-plugin, when the test rebuild runs
[12:02] <cjwatson> doko: pass
[12:02] <cjwatson> doko: trace paths through grub-installer I guess
[12:06] <cjwatson> Mirv: debian/tests/run-tests.sh needs to do the same thing that debian/rules does - run debian/packagekit-check and if it says "no" then add the --disable-packagekit option to configure
[12:07] <cjwatson> Mirv: actually, TBH it could probably just pass --disable-packagekit - the integration tests don't need it
[12:07] <doko> cjwatson, so recursively look at all dependencies for a grub dependency?
[12:07] <cjwatson> Mirv: (unconditionally, I mean)
[12:08] <cjwatson> doko: no, I mean the code.
[12:08] <doko> ahh
[12:08] <cjwatson> doko: grub-installer is the d-i component that installs the actual grub package needed for the system
[12:08] <cjwatson> it probably still has some paths where it installs grub or grub-legacy, rather than some grub2ish grub-whatever package
[12:08] <cjwatson> so if you want to remove grub you need to work out whether those paths are still relevant
[12:09] <cjwatson> and you probably also need to work out what to do about upgrades which I think we may have punted on in Ubuntu
[12:09] <cjwatson> you cannot just say "no dependencies left, it doesn't matter"
[12:09] <doko> ok
[12:14] <jtaylor> hm wasn'T clang in xenial patched to support gccs abi tags?
[12:14] <jtaylor> still get linking errors with boost
[12:27] <showaz> jtaylor: ? http://apt.llvm.org/
[12:32] <Mirv> cjwatson: ok, thanks
[12:35] <jtaylor> showaz: I'm refering to the 1:3.8-2ubuntu4 update
[12:38] <showaz> ubuntu too old upstream software
[12:40] <doko> showaz, there is no newer release
[12:41] <showaz> 3.9
[12:41] <showaz> 4.0
[12:42] <doko> no, there is no release
[12:43] <showaz> 4.9-prelease/rc
[12:43] <showaz> 3.9*
[12:57] <ginggs> jtaylor: llvm 3.8 in xenial is still not quite right, see LP: #1610957 - i'm currently using a backport of 3.81 in yakkety
[13:00] <showaz> ubuntu-llvm used "ninja" for build/deploy?
[13:30] <cyphermox> doko: cjwatson: the only way you install grub (legacy) with grub-installer is if you were about to install grub-pc and someone preseeded grub-installer/grub2_instead_of_grub_legacy=false
[15:07] <seb128> hum
[15:07] <seb128> is pad.ubuntu.com working for others?
[15:08] <seb128> it gives me a "Failed to discover an OpenID server"
[15:08] <Laney> there is something going on
[15:08] <Pici> looks like launchpad is down.
[15:08] <Laney> I just got pagerduty calling me
[15:09] <seb128> launchpad wfm
[15:09] <Laney> about some kind of socket timeout
[15:09]  * Laney screams
[15:09] <seb128> Laney, is #is the right channel for that?
[15:10] <seb128> or is there a #is-something for services
[15:10] <Laney> seb128: they should know I guess
[15:10] <seb128> I asked there
[15:10] <seb128> let's see
[15:12] <Laney> seb128: #is-outage has some activity, guessing it's known
[15:21] <juliank> APT master has completed the switch to CMake. I specifically wrote the translation part so that we generate per-domain templates with headers now (build/po/libapt-pkg5.0.pot, build/po/apt.po, etc), so Launchpad can import them correctly. I hope it works...
[15:21] <juliank> In case you're wondering: Yes, launchpad has not imported apt translations for quite some time.
[15:22] <juliank> This should mean that langpacks can ship updated translations for apt again
[15:23] <seb128> juliank, great ;-)
[15:24] <cjwatson> LP> there was/is a switchover to a new firewall system which seems to have gone a bit wrong, not sure yet whether it's been totally fixed / rolled back
[15:48] <solarce> any chance anyone here has a contact with whomever at Canonical is involved in building and
[15:48] <solarce>  publishing the Ubuntu GCE images?
[15:56] <cyphermox> solarce: what problem are you seeing? maybe we can solve it right here, otherwise we can get you in contact with the right people
[16:25] <solarce> cyphermox: it seems like the setting to disable ipv6 by default in the GCE VMs was lost in an update to the images, at least for Trusty. But I'm not sure which because we just went from a 2015 release toa 2016 release yesterday
[16:25] <solarce> cyphermox: we were based on ubuntu-1404-trusty-v20150909a before and now we're based on ubuntu-1404-trusty-v20160627
[16:26] <solarce> cyphermox: https://github.com/travis-ci/packer-templates/commit/94f32ea4a6bb8c56c9dd3677a5c4708678b6ae8b#diff-8ed6e32f5d271c8bce697243861ea28aL31
[16:26] <solarce> cyphermox: we (travisci) because our build environment GCE images off the official Ubuntu ones, we do a packer + chef build process to customize it
[16:31] <solarce> cyphermox: is the tooling that's used to build the images for GCE public?
[16:58] <cyphermox> solarce: ok
[16:59] <solarce> i've been digging around, http://blog.utlemming.org/2014/11/gce-anyone.html indicates three people to reach out to in #ubuntu-server, so i've taken my question there
[17:10] <juliank> solarce: Did you guys also notice some network/port issues on trusty travis? The apt tests cannot seem to connect to ports on localhost they opened before. I'm not entirely sure what the cause of that is, but it used to work until yesterday
[17:11] <solarce> juliank: yes, it seems to be a side effect of ipv6 being enabled
[17:11] <solarce> juliank: we're building new trusty images with ipv6 disabled via sysctl.conf and should have that live in an hour or two
[17:12] <juliank> solarce: Great.
[17:12] <solarce> it took a long time to notice because most of the user bug reports where things like selenium failing to bind a port, etc
[17:12] <solarce> life in the cloud
[18:12] <solarce> cyphermox: thanks for the help
[18:13] <cyphermox> solarce: yeah, you'd want Odd_Bloke to respond to questions I guess
[18:13] <solarce> we got in touch in #ubuntu-server
[18:15] <cyphermox> ok good :)
[23:23] <infinity> GunnarHj: You around?
[23:24] <Unit193> I am.
[23:24] <GunnarHj> infinity: Yep
[23:28] <infinity> GunnarHj: Looks like https://sourceware.org/bugzilla/show_bug.cgi?id=9809 (and the patch we carry for it) needs updating for 2.24
[23:29] <infinity> GunnarHj: That's just the first failure I hit, I suspect most of our out-of-tree locales will need updating (because of Unicode 8.0)
[23:29] <infinity> GunnarHj: See, eg https://launchpad.net/~adconrad/+archive/ubuntu/ppa/+build/10596201
[23:30] <infinity> GunnarHj: Side-note, I really wish we didn't have out-of-tree locales. :P
[23:30] <infinity> (Backported ones from trunk, sure, but not random locales from wherever)
[23:38] <GunnarHj> infinity: I'll take a closer look. Are you saying they fail to build in next glibc version? As regards new locales, once in a while it's justified to be fast for motivational purposes. When a group of people wants to establish a new language, they shouldn't need to wait for months/years IMO. (But those situtations don't happen very often.)
[23:54] <GunnarHj> infinity: Will compare the patches which bring additional locales, and compare them with one of the 'established' ones. Getting back to you.