[15:00] <barry> o/
[15:00] <tdaitx> \o
[15:01] <cyphermox> o/
[15:01] <robru> o-
[15:01] <sil2100> \o
[15:01]  * slangasek waves
[15:02] <slangasek> #startmeeting
[15:02] <meetingology> Meeting started Thu Aug  4 15:02:34 2016 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
[15:02] <meetingology> Available commands: action commands idea info link nick
[15:02] <slangasek> [TOPIC] Lightning round
[15:02] <slangasek> $ echo $(shuf -e barry doko bdmurray slangasek caribou infinity sil2100 robru cyphermox pitti tdaitx xnox chiluk mwhudson)
[15:03] <slangasek> bdmurray pitti mwhudson robru chiluk barry cyphermox infinity xnox slangasek doko caribou tdaitx sil2100
[15:03] <slangasek> and... go
[15:03] <sil2100> Phew, will have time to prep my notes at least
[15:03] <bdmurray> irc discussion regarding the haproxy charm, switched one being used
[15:03] <bdmurray> worked on utilizing swift for ET code storage in mojo spec
[15:03] <bdmurray> worked on RT re Lift and shift error-tracker staging environment
[15:03] <bdmurray> updated meta-release files for X point release and W EOL
[15:04] <bdmurray> ubuntu-release-upgrader bug triage
[15:04] <bdmurray> reported, fixed update-manager bug regarding hwe-support-status (LP: #1607983)
[15:04] <bdmurray> built, tested, uploaded update-manager w/ HWE support for Trusty
[15:04] <bdmurray> ✔ done
[15:04] <bdmurray> research into adding update-notifier hwe support for Trusty
[15:05] <slangasek> pitti is off
[15:05] <slangasek> mwhudson is in the shadow of the sun
[15:05] <slangasek> robru:
[15:05] <robru> ah
[15:05] <robru> lp:bileto
[15:05] <robru> * big cache expiry cleanup. better logging, more aggressive purging, purging some caches that were previously overlooked and growing without bound
[15:05] <robru> * pretty display for merge targets when there are too many
[15:05] <robru> * fix race condition when deciding when to branch and when to pull local bzr caches
[15:05] <robru> * fix discovery of arches to consider for regressions when overlay PPA supports fewer arches than distro
[15:05] <robru> * serialize some overly-parallelized 'bzr missing' invocations that were hammering LP too hard
[15:05] <robru> * decrease status updater from 15 minutes to 20 minutes also due to too-high load
[15:05] <robru> * new page that only shows failure logs now that status job is spamming the "all logs" page
[15:05] <robru> * make britney skip QA-approved tickets in addition to the already skipped britney-approved tickets, halving britney run time from 1hr to 30 minutes.
[15:05] <robru> * fix race condition where status job reports stale status too soon after build job triggered, so people stop getting confused and building twice in a row for no reason
[15:05] <robru> * change instance_id from uuid to md5sum of html/js/css assets so forced page reloads stop happening when assets haven't changed
[15:05] <robru> * several iterations on CSS based on user feedback; now status/signoffs have independant coloring, and icons, rather than trying to choose one color based on the whole ticket status
[15:06]  * xnox shakes fist at the sun shadow
[15:07] <slangasek> robru: cache expiry cleanup> so things that were being cached that should have been cleaned up when a ticket was closed out (landed/dropped) but weren't?
[15:07] <chiluk> lp 1607920 : zfs failure on first boot due to missing /etc/mtab.  Will be fixing upstream to use /proc/self/mounts, and then SRU back into Ubuntu.
[15:07] <chiluk> Vacation Monday, Tuesday, off again Friday (lasik)
[15:07] <chiluk> -done-
[15:07] <barry> ubuntu-image: gadget.yaml branch; snap weld branch; test fixing branches; more-parsing branch; LP: #1605434
[15:07] <slangasek> "missing /etc/mtab" pff ancient software
[15:07] <barry> debuntu: voluptuous 0.9.2-1, 9.3-1, syncpackage; xonsh 0.4.4+dfsg-1 sponsorship; zope.interface 4.2.0-2
[15:07] <barry> --done--
[15:08] <chiluk> yeah.. it should've been fixed long ago.
[15:08] <robru> slangasek: yeah some lplib cache files were being kept forever, that got up to 200MBs until I started pruning it, also it was keeping britney indexes forever eg so we had trusty index files despite the fact that no trusty tickets were using britney for a while, so I started cleaning those up
[15:08] <chiluk> but there's a lot of logic around it and processing it.
[15:08] <slangasek> robru: oh interesting, I thought lplib did its own cache management properly
[15:08] <slangasek> well, I mean, I wouldn't be surprised if I have a 200MB cache here :)
[15:09] <robru> slangasek: yeah that was my assumption too but I've been watching the lplib cache expand for a long time, it started at 2MBs and just got bigger and bigger
[15:09] <robru> slangasek: it was growing by about 2MBs/day, now I delete all files older than 24 hours and the size hovers at 120MB
[15:10] <slangasek> cyphermox:
[15:10] <cyphermox> yup
[15:10] <cyphermox> MIR:
[15:10] <cyphermox> - review barbican (bug LP: #1543754)
[15:10] <cyphermox> - review content-hub (bug LP: #1597453)
[15:10] <sil2100> Didn't know lplib was such a beast
[15:10] <cyphermox> - review libpod-constants-perl
[15:10] <cyphermox> - review network-manager-openvpn (bug LP: #1574576)
[15:10] <cyphermox> trusty:
[15:10] <cyphermox> - NVMe SRU for ubiquity installer (bug LP: #1600308)
[15:10] <cyphermox> xenial:
[15:10]  * bdmurray thought everybody knew
[15:10] <robru> sil2100: bileto uses it rather aggressively
[15:10] <cyphermox> - grub2 efi ipv6 patch backports (bug LP: #1229458)
[15:10] <cyphermox> yakkety:
[15:10] <cyphermox> - merge / review choose-mirror
[15:10] <slangasek> yay phone MIRs \o/
[15:10] <cyphermox> - packagekit 1.1.1 no-change rebuilds
[15:10] <cyphermox> - discussion/planning shim work with pjones
[15:10] <cyphermox> - review ubiquity gtk3.20 changes from Laney
[15:10] <cyphermox> - fixup ubiquity ftbfs and error in nvme regex
[15:11] <cyphermox> - investigated porting click to packagekit 1.1.1
[15:11] <cyphermox>   - looks like maybe just a no-change rebuild will do, have yet to test from a PPA package
[15:11] <robru> my local lplib cache is only 18MBs, heh
[15:11] <cyphermox> other stuff:
[15:11] <cyphermox> - setting up snap build for canonical-foundations
[15:11] <cyphermox> - shim Secure Boot prompting improvements (bug LP: #1595611, bug LP: #1604936)
[15:11] <cyphermox> - some NetworkManager DNS issue with VPNs not in split-tunnelling.
[15:11] <cyphermox> - helped out slashd with packaging issues (mostly answering questions)
[15:11] <cyphermox> - waiting for a shim signature
[15:11] <cyphermox> yeah, phone MIRs, I'm a bit behind in reviewing MIRs so I took a bite from my list :)
[15:11] <cyphermox> (done)
[15:12]  * sil2100 gently pokes infinity 
[15:12] <slangasek> cyphermox: don't know if it's been mentioned anywhere to you, but the goal from the client side is to have the unity8 stack on the desktop image as a non-default session by 16.10, so there'll be a few more MIRs coming :)
[15:13] <cyphermox> slangasek: I know! no problem, I don't mind reviewing them
[15:13] <slangasek> right :)
[15:13] <sil2100> We moving unity8 stuff to main now?
[15:13] <slangasek> sil2100: yes
[15:13] <cyphermox> it takes a bit of time, maybe I'm a little more thorough than I should by doing some code review, but hey
[15:13] <sil2100> Wow, didn't hear there was a final decision to do that now finally
[15:13] <sil2100> I might want to help with that
[15:13] <slangasek> long overdue, and the security team has been looped in to triage things globally wrt what is going to need deep security review and when
[15:14]  * slangasek pokes infinity less gently than sil2100 
[15:14] <cyphermox> he wasn't responding earlier, but he pinged me at some insane hour
[15:14] <xnox> ..
[15:14] <xnox> boost1.61 update prepared for gcc+icu transition
[15:14] <xnox> openssl enable asm and apply fixes lp #1601836 also requested the same to be done in debian (maintainers are agreeable to the idea)
[15:14] <xnox> generate cloud images appx for Ubuntu on Windows
[15:14] <xnox> 14.04 point release testing (nvme)
[15:14] <xnox> sru validation
[15:14] <xnox> upcoming helping with boost/gcc/icu transition
[15:14] <xnox> ..
[15:15] <xnox> .... clearly enabling asm optimisations never causes regressions =) *caugh* but testing so far is solid and everyone else has it enabled
[15:16] <slangasek> xnox: was the '..' a 'done'?
[15:16] <cyphermox> isn't it an escape in some weird terminal thing?
[15:16] <slangasek> xnox, doko: I noticed the thread discussing ABI incompatibilities between gcc5 and gcc6... I think we're going to need to dig into this wrt phone stack ABIs
[15:17] <slangasek> (not here and now, but imminently)
[15:17] <slangasek> cyphermox: 8270?
[15:17] <slangasek>  * last week: CDO midcycle sprint
[15:17] <slangasek>   * lightning talk demo of snappy first boot experience, it was pretty snappy!
[15:17] <slangasek>   * discussions around snappy support on 14.04
[15:17] <slangasek>  * Monday: travel swap
[15:17] <slangasek>  * Tuesday: email catch-up
[15:17] <slangasek>  * Wednesday: sick day
[15:17] <slangasek>  * next:
[15:17] <slangasek>   * continued mid-cycle braindump
[15:17] <cyphermox> slangasek: sorry, I 404.
[15:17] <slangasek>   * continued email catch-up
[15:17] <slangasek>   * ubuntu-image
[15:17] <slangasek> (done)
[15:17] <xnox> slangasek, we know boost-regexp is bust, and autopkgtest shows that possibly some kdelibs are bust too. Apperantly they have fixed the bugs in abi-tagging in nested templates or something.
[15:18] <doko> - GCC 6 now the default
[15:18] <doko> - gcc 4.9.4 release, package update, gcc 5.4.1 package update
[15:18] <doko> - boost 1.61 transition
[15:18] <doko> - fixing build failures related to GCC 6 and boost 1.61
[15:18] <doko> - pinging people about fixing build failures ...
[15:18] <doko> - openjdk-8 update
[15:18] <doko> (done)
[15:18] <slangasek> xnox: sure; but where "fixed bug" == "break compatibility", we need to have a plan :)
[15:18] <doko> slangasek, but it will be ugly in any case
[15:18] <slangasek> it sounds like the gcc6 + boost 1.61 build failures should now be fixed en masse? yes?
[15:19] <doko> yes, there are "some".
[15:19] <doko> but let's first finish the transition rebuilds, see http://people.canonical.com/~ubuntu-archive/transitions/html/boost1.61.html
[15:19]  * slangasek nods
[15:20] <slangasek> doko: so I suppose once boost has cleared we want to revisit what's still broken that the phone team needs to look at?
[15:20] <slangasek> meanwhile
[15:20] <slangasek> no caribou
[15:20] <slangasek> tdaitx:
[15:20] <tdaitx> = OpenJDK security update
[15:20] <tdaitx> - Packaged IcedTea 2.6.7 into openjdk-7-7u111-2.6.7 for Trusty and Precise; tested and provided packages to the security team
[15:20] <tdaitx> * Had to retry ARM64 build on Trusty 6 times until gcc/g++ didn't segfault (happens on every update as the PPA must use only the -security pocket, version on -updates works fine); tips on how to fix that?
[15:20] <tdaitx> - Working on 2 regressions from the security updates:
[15:20] <tdaitx> * JAMVM regression on OpenJDK 7u111
[15:20] <tdaitx> * OpenJDK 8u101 Zero VM on armhf (LP: #1609190)
[15:20] <tdaitx> = AOB
[15:20] <tdaitx> - Trying to figure out why Network Manager can't restart openvpn when ping-restart is set (seems to be missing the up script or dependencies in the chroot, still debugging that, see https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=667611#20); cyphermox, help? =)
[15:20] <doko> slangasek, yes, they have a list, and are apparently starting to fix things
[15:20] <tdaitx> - Provider will install landline/adsl tomorrow morning (Friday)
[15:20] <tdaitx> (done)
[15:21] <slangasek> doko: ok good :)
[15:21] <cyphermox> oh, hai
[15:22] <sil2100> oh
[15:22] <sil2100> Me
[15:22] <sil2100> - Landing team work, silo coordination, preparing landing e-mails
[15:22] <sil2100> - RTM status meetings
[15:22] <sil2100> - Touch xenial:
[15:22] <sil2100>   * Coordinating rebuilds of all core clicks for arm64
[15:22] <sil2100>   * Modifying the livecd-rootfs build hooks to fetch arm64 from a different location
[15:22] <slangasek> tdaitx: tips on how to fix the -security gcc being broken> get the security team to pull that gcc into the security pocket for you? :)
[15:22] <sil2100>   * Switching system-image to import the new arm64 rootfs - first frieza arm64 image!
[15:22] <sil2100>   * Discussing fat-click-package future (consensus: wait for snaps?)
[15:22] <sil2100> - DMB meeting
[15:22] <sil2100> - Trying to refresh packagesets with the packageset script, output strange
[15:22] <sil2100> - Added mate-hub to the ubuntu-mate packageset
[15:22] <sil2100> - Releasing the modified goget-ubuntu-touch for new device support
[15:22] <sil2100> - Looking into backporting libboost1.58 to the overlay as per phone needs
[15:22] <sil2100> - Staging updated touch seeds to include new packages (indicator-keyboard, account-plugin-owncloud)
[15:22] <sil2100> - Started backporting golang-1.6 to the overlay as it seems to fix various issues
[15:22] <sil2100> - Finished work on switching generate-commitlog to the new bileto published_versions
[15:22] <sil2100> - Validating SRUs for xorg-server and mesa
[15:22] <sil2100> (got distracted, done)
[15:22] <tdaitx> slangasek, heh, I will poke them then
[15:22] <tdaitx> thanks
[15:23] <slangasek> tdaitx: they may say 'no', of course, but it's the right conversation to have :)
[15:23] <slangasek> sil2100: an arm64 image? woot!
[15:24] <sil2100> \o/
[15:24] <slangasek> sounds like you all had a very productive week without me
[15:24] <sil2100> Yeah, it doesn't boot quite yet, but it's because we didn't have a working device tarball
[15:24] <slangasek> should I go away more often? ;)
[15:24] <slangasek> any questions re: status?
[15:24] <sil2100> ondra assumed the armhf one should work ;p
[15:24] <sil2100> Not from me
[15:25] <slangasek> sil2100: backporting golang-1.6 to the overlay> that seems like it shouldn't be necessary to "backport" given that we have golang-1.6 as an SRU all the way back to trusty.  Do we just need to forward-copy to vivid/wily?
[15:26] <sil2100> Oh, might be a good idea then, yeah
[15:26] <sil2100> Didn't know we had that in trusty
[15:26]  * sil2100 could have checked
[15:26] <sil2100> I mean, I rarely look 'backwards' for stuff ;)
[15:26] <sil2100> That would make things rather easy
[15:27] <slangasek> if we want to do it "properly" then we may have to re-bootstrap in each release; mwhudson would be the one to talk to about that
[15:28] <slangasek> otherwise, I would be ok in this case with just binary-copying forward from trusty to vivid+wily, provided that infinity won't stab me for this
[15:28] <sil2100> He's not here so I guess we can just do it, he won't notice ;p
[15:29] <slangasek> heh
[15:29] <slangasek> sil2100: anyway, fast path to avoid blocking on SRU process would be for you to just do binary copy from trusty to the overlay ppa; it's a completely new package so I *think* it should be safe to copy in without going through a silo
[15:30] <slangasek> but, of course, you should double-check that it's not going to clobber any other binary packages
[15:30] <sil2100> I wanted to do that through a silo so that I can do rebuilds of all the packages we need and see if those work ok on vivid
[15:30] <slangasek> ok
[15:30] <slangasek> then you can binary-copy to a silo instead of to the overlay :-)
[15:30] <sil2100> SInce we know that golang-1.6 fixed some issues for xenial, and possibly it will fix those on vivid too
[15:30] <sil2100> Will do, thanks :)
[15:31] <slangasek> are you intending to change the golang defaults, or just provide golang-1.6 as an alternative?
[15:31] <slangasek> the latter matches what's been done in trusty, and is probably easier to QA in pieces because only things you know are broken and need rebuilt have to opt-in
[15:31] <sil2100> Not sure yet, both ways would work fine as we want to use the new golang for all our touch go packages anyway
[15:32] <sil2100> But yeah, maybe it would be less troublesome to not do a change of defaults
[15:32]  * slangasek nods
[15:32] <sil2100> THen we can have more control if anything is broken
[15:32] <sil2100> +1 on that idea then ;)
[15:32] <slangasek> cool
[15:32] <slangasek> [TOPIC] AOB
[15:33] <slangasek> anything else?
[15:34] <slangasek> #endmeeting
[15:34] <meetingology> Meeting ended Thu Aug  4 15:34:56 2016 UTC.
[15:34] <meetingology> Minutes:        http://ubottu.com/meetingology/logs/ubuntu-meeting/2016/ubuntu-meeting.2016-08-04-15.02.moin.txt
[15:34] <slangasek> thanks, all!
[15:35] <barry> thanks!
[15:35] <sil2100> Thanks!
[20:06] <mwhudson> slangasek, sil2100: bootstrapping golang-1.6 in vivid and wily should be easy enough, gccgo is new enough there
[20:06] <mwhudson> slangasek, sil2100: copy-package -b is even easier though ...
[20:43] <sil2100> mwhudson: I did a copy-package to a silo already from trusty
[20:44] <sil2100> mwhudson: it's now about building go packages
[20:44] <sil2100> Like, those using golang
[20:44] <mwhudson> ah ok