[06:11] <osama> hi
[15:03]  * slangasek waves
[15:03] <sil2100> o/
[15:03] <jodh> \o
[15:03] <mvo_> hi
[15:04] <slangasek> #startmeeting
[15:04] <slangasek> hmm where are those bots!
[15:04] <slangasek> (anyone know where meetingology has gone?)
[15:05] <slangasek> [TOPIC] Lightning round
[15:05] <sil2100> huh
[15:05] <slangasek> $ echo $(shuf -e barry doko stgraber jodh bdmurray slangasek cjwatson xnox caribou infinity mvo bhuey sil2100 robru)
[15:05] <slangasek> bdmurray caribou infinity bhuey jodh stgraber xnox slangasek cjwatson sil2100 barry mvo robru doko
[15:06] <slangasek> bdmurray is at a conference today
[15:06] <slangasek> caribou:
[15:06] <caribou> * Work on MAAS packaging issue
[15:06] <caribou> * DMB application for PPU on sosreport and makedumpfile - Awaiting email vote : no Quorum during the meeting
[15:06] <caribou> * Some kernel basic debugging
[15:06] <caribou> (done)
[15:06] <caribou> * Work on apt offline queries with mvo's help
[15:06] <robru> o/
[15:07] <sil2100> robru: off to bed! ;)
[15:08] <slangasek> infinity not here
[15:08] <slangasek> bhuey not here
[15:08] <slangasek> jodh:
[15:09] <jodh> * upstart
[15:09] <jodh>   - bug 1360208:
[15:09] <jodh>     - Simplified code.
[15:09] <jodh>     - Still working on tests.
[15:09] <jodh>   - Fixed logrotate job bug in branch which ensures logrotate runs even
[15:09] <jodh>     if its state file is corrupt.
[15:09] <jodh>   - Testing 1.13.2 on Touch.
[15:09] <jodh>   - Wrote testplan for utopic sync to Touch:
[15:09] <jodh>     https://wiki.ubuntu.com/Process/TestPlans/upstart
[15:09] <jodh> * systemd-shim:
[15:09] <jodh>   - Wrote testplan for utopic sync to Touch:
[15:09] <jodh>     Wrote https://wiki.ubuntu.com/Process/TestPlans/systemd-shim
[15:09] <jodh> * MIR
[15:09] <jodh>   - worked on bug 1349899.
[15:09] <jodh>     - raise debian bugs 760979 + 761177.
[15:09] <jodh>   - sent a few questions+suggestions to the mir-team.
[15:09] <jodh> * cgmanager
[15:09] <jodh>   - Reviewing changes.
[15:09] <jodh> ⩩
[15:11] <slangasek> stgraber:
[15:13] <slangasek> I'll go ahead, and we'll circle back to stgraber
[15:13] <slangasek>  * working on systemd-shim/cgmanager updates for the current systemd APIs, sponsoring into Debian; the Debian bug report has been reopened, hallyn and desrt investigating (Debian bug #756076)
[15:13] <slangasek>   * shepherded fix for bug #1365095 into ubuntu-rtm
[15:13] <slangasek>  * gave IBM feedback on debian/copyright file for java in multiverse
[15:13] <slangasek>  * making plans to attend the cloud sprint in November
[15:13] <slangasek>  * working to identify fat that can be trimmed from the phone rootfs - we could use better scripts for this!
[15:13] <slangasek>  * reviewed more TI packages for the partner archive
[15:13] <slangasek>  * still working on filling our open role - if you have good candidates you'd recommend, it's not too late for them to get their name in
[15:13] <slangasek> (done)
[15:14] <stgraber> slangasek: back, network problem...
[15:14] <stgraber>  - Quite a bunch of internal meetings and working on internal projects and
[15:14] <stgraber>    documentation, so not much I can share, sorry.
[15:14] <stgraber>  - LXC code reviews and landing a bunch of patches.
[15:14] <stgraber>  - Did some FFe reviews.
[15:14] <stgraber>  - Update system-image server code to hash all the custom tarballs too.
[15:14] <stgraber>  - Setup some extra custom tarballs on the RTM channels.
[15:14] <stgraber>  - Setup a Windows8 UEFI environment for ubiquity testing (bug 1265192)
[15:14] <stgraber> (DONE)
[15:14] <sil2100> slangasek: you probably should remove xnox from the list of attendees during shuffle ;)
[15:15] <slangasek> sil2100: yeah, I accidentally cut'n'pasted from the wrong log
[15:15] <slangasek> cjohnston:
[15:15] <slangasek> sigh
[15:15] <slangasek> cjwatson:
[15:15] <cjwatson> Coordinated libav11 transition, and cleaned up a few other bits of noise in proposed-migration.
[15:15] <cjwatson> Helped out with getting click signature support re-landed; lots of fighting with CI Train.  Fixed up a loose end in the security tests (lp:~cjwatson/qa-regression-testing/click-install-untrusted).
[15:15] <cjwatson> Landed UbuntuKylin image changes to use their own seeds.
[15:15] <cjwatson> Fixed Kubuntu Plasma image builds (by rebuilding kfilemetadata-kf5 against libav11 in their PPA).
[15:15] <cjwatson> Somewhat caught up on miscellaneous reviews.
[15:15] <cjwatson> Working on Debian #760148 (grub-mount hang); now have a test image from bug submitter, so need to sit down with a debugger and analyse it.
[15:15] <cjwatson> Working on bug 1342858.  I think I have most of a fix, but need to test it locally before proposing a merge.
[15:15] <cjwatson> TODO: Set up MATE image builds; start on putting the pieces together for package acquisition in click (allowing use of ubuntu-download-manager as a plugin, but also with native support for use elsewhere).
[15:15] <cjwatson> ..
[15:16] <sil2100> o/
[15:16] <sil2100> - Annual holidays from 8th till 10th of September - nothing done during that time, so short week
[15:16] <sil2100> - Reading up e-mails
[15:16] <sil2100> - CI Train maintenance and features:
[15:16] <sil2100>   * Finishing sync: enhancements, testing and deployment
[15:16] <sil2100>   * Fixing some PEP8 voilations in existing branches
[15:16] <sil2100>   * Unit tests for sync features
[15:16] <sil2100>   * Start work on some sync: corner-cases
[15:16] <sil2100> - Reading up on PEP8 and pyflakes
[15:16] <sil2100> - Documentation of new features and recommendations
[15:17] <sil2100> - Patch Pilot work:
[15:17] <sil2100>   * Check the libgksu bug-fix release MR, comment on the fixes needed
[15:17] <sil2100>   * Checking libgksu FTBFS - Debian merge needed
[15:17] <sil2100> (done)
[15:17] <barry> phone: LP: #1365646.  LP: #1279970.  phone phlashing phun.  LP: #1365991 (discussions/triage); LP: #1367004; LP: #1283783; LP: #1367014; LP: #1367016 (actually, we won't need to MIR tox); LP: #1367028; LP: #1367041 (also won't need); LP: #1349832 (also won't need).
[15:17] <barry> debuntu: more git-dpm & git packaging experiments.  all results have been posted to debian-python@ and i am recommending git-dpm.  tox 1.7.2-1; core-cov 1.14.0-1 and debian bug #760583; syncpackage: lazr.config, tox, cov-core, lazr.delegates, lazr.smtptest, flufl.bounce
[15:17] <barry> other: python issue 16662 (load_tests in packages)
[15:17] <barry> (done)
[15:18] <mvo_> Busy week, gtimelog gave me a >60 line summary that I think is too
[15:18] <mvo_> boring^Wlong. So the short version (unless someone wants the long
[15:18] <mvo_> one in which case I'm happy to paste it too):
[15:18] <mvo_> click: worked with Colin to land click 0.4.32.1 and the SDK with
[15:18] <mvo_> re-enabled signed click support. This was a long and tricky landing,
[15:18] <mvo_> thanks to Colin for his help with that. I also spend some time on the
[15:18] <mvo_> SDK to push some drive-by fixes for issues I noticed during running
[15:18] <mvo_> the (loooong) SDK testplan. Plus some (multiarch) work on using
[15:18] <mvo_> ubuntu-sdk-libs-dev:$ARCH directly to build the chroot, pushed some
[15:18] <mvo_> branches for this, should be ready now. And some work to identify if a
[15:18] <mvo_> binary in a click package is using libs that are not in the SDK.
[15:18] <mvo_> apt/python-apt: maintenance, some bugfixes and merges
[15:18] <mvo_> Also worked onr software-properties (support for apt-add-repository
[15:18] <mvo_> for ubuntu-rtm), squid-deb-proxy, unattended-upgrades and uploaded new
[15:18] <mvo_> versions.
[15:18] <mvo_> (done)
[15:18] <robru> * Major ongoing overhaul of CI Train:
[15:18] <robru>   - deleted several hundred lines of untested, *untestable* garbage glue code
[15:18] <robru>   - implemented several forms of static code analysis, including pep8, pyflakes, and xml validation in order to detect problems before they get committed to trunk
[15:18] <robru>   - gated MPs against lp:cupstream2distro on successful runs of the test suite
[15:18] <robru>   - ripped out jinja2 (a templating engine) in favor of using python's builtin string.format, with new unit tests to confirm that it works as expected.
[15:18] <robru>   - rewrote "setup-citrain" script almost from scratch, it used to be a weird big pile of spaghetti, with many functions passing many inscrutably-named arguments back and forth all over the place, into a single class with a few scrutibly-named instance attributes
[15:18] <robru>   - many more enormous changes planned, but not yet implemented due to a couple days of illness
[15:18] <robru> ✓
[15:19] <mvo_> cjwatson: I'm curious about the plans for the click acquire, have you written up something somewhere about it?
[15:20] <doko> - openjdk-7 update for utopic and trusty, backports for precise and lucid
[15:20] <doko> - openjdk-8 update for utopic, backports for trusty and precise
[15:20] <doko> - GCC update, Linaro GCC update
[15:20] <doko> - NBS down to zero, looking at build regressions the other time
[15:20] <doko> - started test rebuild for trusty/main
[15:20] <doko> - valgrind update for ppc64el
[15:20] <doko> - MIR mentoring, finished sparql-wrapper-python
[15:20] <doko> - people pestering about the state of the nation^Warchive
[15:20] <doko> (done)
[15:21] <cjwatson> mvo_: it is but a glimmer in my eye
[15:21] <jodh> doko: yes thanks for your help on the MIR!
[15:21] <mvo_> :)
[15:21] <cjwatson> mvo_: though I rather like the apt methods architecture for this and am tempted to steal it
[15:22] <mvo_> cjwatson: yeah, I think that makes sense, we could even use their transports if the protocol is compatible
[15:22] <cjwatson> interesting suggestion, will look into that
[15:22] <cjwatson> (and I don't mean that in the British way)
[15:23] <slangasek> any questions over status?
[15:23] <slangasek> btw, one thing from my status was that I'm trying to identify packages that shouldn't be on the phone
[15:23] <mvo_> cjwatson: ok, please let me know if you want to offload (some of) this work
[15:23] <slangasek> I've thought of what I think would be a useful representation of this
[15:23] <cjwatson> will do
[15:24] <slangasek> and wonder if there's any prior art on it (so I don't go reimplementing things from scratch)
[15:24] <cjwatson> slangasek: did I mention germinate's rdepends output?
[15:24] <cjwatson> though I guess that's per-package
[15:24] <slangasek> cjwatson: ah, I don't think you mentioned it was rdepends I should look at
[15:24] <slangasek> so what I think I want is, for a set of packages in the image, to automatically identify individual dependencies which, if cut, would have the greatest impact on size reduction
[15:25] <cjwatson> rdepends gives you the complete multiply-branched backtracked tree for each package
[15:25] <slangasek> (greatest impact in terms of number of packages or size)
[15:25] <cjwatson> slangasek: I would be very interested in having such a thing in germinate, and it *might* not be that hard
[15:25] <cjwatson> it has most of the computation you need already, I think
[15:25] <cjwatson> though, I'm not sure I'm volunteering to do it
[15:25] <slangasek> ok, I'll look at integrating it there, thanks
[15:25] <mvo_> python-apt can probably simulate that for you too
[15:26] <slangasek> cjwatson: would we want to output this analysis by default?  I'm wary of slowing down germinate runs
[15:27] <cjwatson> could just be controlled with an option; rdepends has such a thing (though on by default)
[15:27] <slangasek> hmm, the other thing is that I care about this at the image level, not necessarily at the seed level
[15:27] <cjwatson> for rdepends the slowness isn't the computation, but the I/O
[15:27] <cjwatson> well, images map pretty closely onto a set of seeds
[15:27] <cjwatson> analysing it for each seed would be a decent enough approximation, and you'd only have a few outputs to look at
[15:28] <mvo_> slangasek: so what kind of input would you give the tool? would you want to run it for each (non-required) package on the image? each leave package? or by user choice?
[15:28] <cjwatson> but if it doesn't fit it doesn't fit :)
[15:28] <slangasek> cjwatson: a set, yes, but not an individual seed... I found trying to walk multiple seeds for the phone analysis annoying (particularly when trying to use the web output, which is the default)
[15:28] <slangasek> anyway, that gives me some ideas... I'll run with it now, thanks :)
[15:29] <cjwatson> yeah, I can see that
[15:29] <slangasek> mvo_: I guess the input should be the set of all packages in the image, and the output would be the list of all packages with 0 or 1 reverse-dependencies on the image ordered by size impact
[15:30] <slangasek> #topic AOB
[15:30] <slangasek> anything else?
[15:30] <cjwatson> ideally I think you want small clusters as well, but I appreciate it gets non-trivial
[15:31] <cjwatson> this feels like it ought to be a standard graph analysis problem
[15:31] <cjwatson> weighted graph rather
[15:31]  * slangasek nods
[15:31] <cjwatson> maybe I should break out the brick^W^WIntroduction to Algorithms
[15:32] <cjwatson> (hm, where did I leave my copy?  none of my bookshelves appear to be bowed in the middle)
[15:32] <barry> ;)
[15:33] <slangasek> sounds like that's it today then
[15:33] <slangasek> #endmeeting
[15:33] <slangasek> thanks all :)
[15:33] <mvo_> thanks
[15:34] <barry> thanks!
[15:34] <caribou> thanks!
[15:35] <jodh> thanks!
[15:35] <stgraber> thanks!
[15:37] <sil2100> o/