Postings by Dmitrijs Ledkovs

Incomplete autopkgtest dependencies installed from test-trigger.


I'm confused by this run of snapd test, as triggered by systemd from proposed.
The one in release pocket, ends in 4 and the one in proposed ends in 5

log url is: <a href="" title=""></a>
artifacts url is:
<a href="" title=""></a>

--apt-pocket=proposed=src:systemd --apt-upgrade snapd

/usr-merge by default for new installations, with backwards compat

Over the past year, there have been multiple times when I had to fix
scripts / exec calls to account for split-usr and usr-merged systems
by patching constants or adding configuration checks to detect if
certain binaries need to be called from /usr or /.

At the same time, I find certain properties of usr-merged systems
attractive. For example, the fact that all the packaged managed stuff
is scoped in /usr is very convenient, for deduplication, sharing,
snapshotting, including/excluding said dir in configs for diffs &

The improved 18.04.1 LTS Server Installer - Call for testing!

Download from: <a href="" title=""></a>

With the release of 18.04 LTS Bionic Beaver the new server installer
was introduced. At the time, it still lacked certain critical features
which have now been implemented.

For 18.04.1 we are adding support for LVM, RAID, VLAN, Bonds to the
new installer. The functionality and UI are all in place. It has not
yet been refined, thus I expect minor UI changes or improvements to
these features for Cosmic release.

ubiquity migrated to git

$ git clone lp:ubiquity
$ git push lp:~user/ubiquity

is all you need to know if you already have the git lp: url setup in
your ~/.gitconfig
[url "git+ssh:// ... at git dot"]
insteadof = lp:

(See <a href="" title=""></a> for more details)

lp:ubiquity is owned by ~ubuntu-installer team, and thus access rights
should be same as the bzr branches.

This repository has most of the tags, and currently has three branches:
* xenial
* bionic
* master

I expect master branch to be used for the currently-in-development
release, and branches created on the as-needed basis for

Road to new openssl

openssl has changed api/abi. Currently Ubuntu ships 1.0.2 LTS series
openssl. Newer api/abi is available as a non-lts 1.1.0 series. Both
1.0.2 and 1.1.0 series will go end of life upstream over the lifetime
of bionic.

TLS 1.3 is currently undergoing standardisation
(<a href="" title=""></a>) But it seems like it is still
being actively iterated on.

The next openssl series is expected to be 1.1.1 and it should be
binary compatible with 1.1.0 series.

Re: Get involved -> Development page is down


On 2 October 2017 at 21:58, Denis Mekhanikov < ... at gmail dot com> wrote:
I believe that sub-domain is now decommissioned, maybe a redirect
should be put in place to drop people to <a href="" title=""></a>. Where did you
find that url?

Change of scope and target market for i386

NB! this is a mailing list for developers, and this is a _proposal_
that I want to discuss with the *buntu developers.

Removal of upstart and cgmanager from Artful Aardvark

Upstart and CGManager have been in maintenance mode for a few releases
now. The primary users of upstart & CGManager were Mobile (Unity
Touch) and Unity8 desktop sessions. Upstart is used in those session
to provide user session supervisor and implement application

Currently, ubuntu-app-launch has been ported to provide application
lifecycle using systemd. However, neither of the two session packages
were ported to kick off touch/unity8 sessions exclusively under

partman, recipes, sizing, swap, and all the things

Our classic installers support dynamic sizing relative RAM size by
using percent specification. Most commonly this is used for defining
swap, e.g. 200% for the swap of double the size of ram.

However, nowadays the landscape and purpose of the swap have changed
somewhat. Which in the common case results in swap overallocation.
Specifically, for a long time ubuntu has hibernation disabled by
default and encouraging using only suspend. The common RAM sizes of
machines have gone up significantly. And many systems are configured
with very high RAM-to-disk ratios.

/usr/bin/gpg to be provided by 2.1.x series in Yakkety


Recently Debian has switched to 2.1.x by default
<a href="" title=""></a>

And most of the bugs have been fixed by now in the archive, and these
have now made their way to yakkety.

I have requested an FFe which was granted to do the switch in Ubuntu

Installation Media and supportability of i386 in 18.04 LTS Re: Ubuntu Desktop on i386

Hello Bryan,

Let me resurrect this thread. In the context of what we should be
doing in 18.04 and what to do between now and then.

In 2018:
- it will be over 2 years since 3rd party ISVs stopped supporting
software on i386, or even never had it officially
- e.g.

libseccomp 2.3.1 uses negative (pseudo) syscall numbers by default


New libseccomp is in yakkety proposed. There is a change, on some
architecutres, w.r.t. to canonical representation of syscall

There are normal syscall numbers and multiplexed ones. And some are
exposed as both - direct numbers and negative pseudo syscall numbers.
All filtering should remain in place for both direct and pseudo

But I had to adjust our autopkgtests for this, and I'm wondering if
there are any other pieces of software to fix as a result of this
upstream change on some architectures (e.g.

Archive Reorg Episode VII: Follow Build-Depends

tl;dr - xnox wants to remove 1 344 (35%) source packages from main

Google Doc for suggestions & comments:
<a href="" title=""></a>

Sample old/new germinate output is at:
<a href="" title=""></a>

= Archive Reorg Episode VII: Follow Build-Depends =

== Introduction ==

Loosely this builds up on the existing portions of the Archive Reorg
evolution. To recap, we have 4 components in the archive, which are
defined by their licensing on one axis, and based on seeds for the
other axis. Thus e.g.

Description: add ifenslave and vlan, needed by MAAS, required by Ante


iproute2 suite of utilities, and specifically the ip command can do
vlans and network bonding.

Can MAAS use the modern / new commands?
Why are legacy commands being pulled onto the cloud-image?

iproute2 is in minimal, and thus guaranteed to exist even in the
smallest environments.

An excellent guide as to what iproute2 is capable of is available here:
<a href="" title=""></a>

ip link add $master type bond

ip link set $slave master $master

ip link add name eth0.110 link eth0 type vlan id 110

I would like to revert below commit.


Ubuntu Desktop on i386


Ubuntu has an i386 port which is fully supported.

There a bunch of 3rd party applications that rely on the Multi-Arch
technology to support/use i386 binaries on amd64 (e.g. Skype from the
partner archive). BTW, can we ask Microsoft to publish native amd64
binaries, rather than those that rely on multi-arch i386? Also, does
Valve Steam product rely on i386 multiarch binaries? or is it fully
amd64? (and e.g. downloads/bundles/ships any required i386 binaries
that it needs)?

gnupg 2.1.x by default

A growing number of features are desired, yet missing from classic
gnupg (1.4.x series). Specifically support for many gpg smarcards &
tokens, as well as ECC.

I'd like to propose to switch to gnupg 2.1.x by default.

First it would mean upgrading gnupg 2.0.x to 2.1.x. (available in experimental).

Second we need either transition, or actually patching and fixing a
few packages. Most of the patches and transitions are trivial, as
actually usage and output from gpg2 is the same as with gpg.

systemd local bridge

So I've started converting a few things and it seems to work fine.

BindsTo=android- ... at proerty dot name\

WantedBy=android- ... at proerty dot name\

However this cannot deal with "*" / globbing matching as seen in:
vivid/etc/init/android-tools-adbd.conf persist.sys.usb.config=*adb
vivid/etc/init/tethering.conf persist.sys.usb.config=rndis*

Are the glob values actually known there?

upstart-local-bridge plan to support systemd

At the moment upstart-local-bridge is used on the phone in very
specific ways. Android property watcher is run inside the
hardware-support lxc container, and it communicates property changes
to the upstart-local-bridge which makes those changes available to the
system and session inits.

The way property watcher operates will remain the same. However how
the events are exported will need to be modified slightly.

At the moment upstart-local-bridge emits a system upstart event. A few
jobs depend on those.

Please retire

At the moment do-release-upgrade -d from utopic -> vivid is failing
because repository does not exist for vivid, it's
failing to fetch and thus upgrade is aborted becase of "possible
network problems".

Can we please stop adding by default? Not consider
it's absence as failure to upgrade Ubuntu?

gnome-python -> universe

Can we demote gnome-python to universe?

As far as I can tell the only reverse dependency chain is:
compiz -> depends: compiz-gnome -> recommends: python-gconf

Can we port compiz-gnome to using gsettings or something? Is it
actually, in fact, use python-gconf?

Outstanding systemd jobs

I'm going to advertise lp:~upstart-devel/upstart/upstart-jobs again,
which is a branch that stores the result of scanning all init.d
scripts, upstart jobs and systemd units in ubuntu devel and debian

I've also now run a simple (dumb) script to generate a list of upstart
jobs which do not have neither systemd units nor initscripts.

GnuTLS 2.8 transition for Unicorns

GMP in unicorn is now dual licensed under LGPLv3+ & GPLv2+ [1], thus
software only needs to be compatible with GPLv2 to link against

Should we transition to gnutls28 by default?
Should we only have gnutls28 in main?

At the moment we have both gnutls26 and gnutls28 in main.

[1] <a href="" title=""></a>

Odd ADT test results

A few ADT tests are red in jenkins:
* bedtools
* epsilon
* ipset

What happens is that newest version is asked to be tested (e.g.
2.19.1-1 src+bin) yet jenkins sees an older version (e.g. 2.17.0-1
src) which doesn't have any tests and thus marks the job as failed.

Imho, job should be marked unstable and auto-retried.

Since this is a source package missmatch -> how come britney sees both
src package and binaries already yet jenkins does not? Are the two not
operating against the same internal mirror?

Unreproducible test results, blocking python2 removal - Re: [Merge] lp:~xnox/gallery-app/fix-sample-dir into lp:gallery-app

I made a merge proposal to resolve compatibility in gallery_app with
next phablet-tools (which is needed to support transition to python3
and python2 supported branches).

<a href="" title=""></a>

A jenkins bot, points that .debs were built for this project and tests
executed on the device.

However, gallery_app is now a click and tests pass when it's build and
tested as a click.

I've fetched the two generated debs (gallery-app and
gallery-app-autopilot) pushed to my mako device, and executed the
tests using "phablet-test-run -v gallery_

Dmraid to Mdadm migration

This cycle a few things were done in mdadm & dmraid packages to
prepare for transition from dmraid to mdadm for a few formats that
mdadm supports.

At the moment it trusty, mdadm has full support to assemble Intel
Matrix Raid and DDF fakeraid arrays, instead of dmraid. At the moment
however, mdadm codepath is disable via a kernel cmdline options
(nomdmonddf nomdmonisw) and thus if both mdadm and dmraid are present
in the initramfs dmraid will complete assembly. With or without kernel
options, dmraid completes assembly if mdadm is not present.

Shutdown / Logout improvements - Call for testing

As part of <a href="" title=""></a>

recently a number of uploads were done to improve shutdown ordering,
such that there is a smoother transition between lightdm shutting down
and plymouth shutdown splash animation appearing. Ideally, upon
shutting down, one shouldn't see intermittently any console messages.

If you in the past, upon shutdown, used to see some white messages on
black background between desktop session tear-down and plymouth
shutdown splash appearing, please consider testing latest trusty
packages of upstart, lightdm and plymouth.

Cross-compile with CMake from SDK Apps to Unity8/Mir

As of latest cmake upload into trusty, it is now trivial to
cross-compile CMake based projects:

One time setup:
$ mk-sbuild --target armhf trusty

Building a package:
$ sbuild -A -d trusty --host armhf package*.dsc

This succeeds for a wide variety of packages:
* mir
* unity8
* ubuntu-system-settings
* ubuntu-settings-components
* unity-notifications
* unity-scope-onlinemusic
* uhub
* uchardet
* and many more...

Since this is built into CMake, one can use cmake direct:
dpkg-architecture -aarmhf cmake ../; make

(given appropriate build-dependencies installed)

With chroots managed by click: