Postings by Harald Sitter

deleting all kubuntu branches on debian repos


I've just noticed that we still have a gazillion kubuntu_* branches in
the repos on, seeing as they aren't used anymore I
wonder if there are any objections to mass deleting the lot of them?
They serve no purpose other than confuse people.


falcon has no bugzilla product

I literally can't file bug reports :(

releaseme now relying on kdoctools for documentation installation


As per [1] the release management suite releaseme creates tarballs
which rely on kdoctools to process a directory with docbooks inside.
Previously releaseme manually constructed CMakeLists suitable for the
documentation lineup at hand. This required an excessive amount of
fairly poorly engineered code though, so this change should increase
the overall reliability of documentation installation.

Unfortunately, this also comes with a slight downside as previously
releaseme would try very hard to retain logic in the english
documentation (e.g.

anyone interested in working on a KIO slave on libcloudstorage?


I've just been talking to Jean-Baptiste Kempf from VideoLAN about
their new cloud library libcloudstorage [1] which is apparently fairly
usable now and supports a whole bunch of main stream file storage
We are thinking that a KIO slave based on this could be really awesome
and give us access to clouds without having to deal with each cloud

projects api alpha

I finally got annoyed with the obscenely large xml file we had for
projects and wrote a RESTful api to replace it.

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

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

I'll probably finalize it next week unless someone has handy
suggestions for improvement.


Welcoming Rik Mills as Kubuntu Developer

With all present devs voting in favor, Rik got voted into the circle
of Kubuntu Developers. He's done lots of awesome stuff in the past and
we are looking forward to seeing him put these new found powers to
good use.

Congratulations and welcome!


broken logo artwork thing

knome on IRC pointed the following out to me

Screenshot hosting repo (e.g. for Appstream)


We have just finished sorting out a centralized screenshot repo. This
repo should be used for promotional screenshots of KDE software.
Specifically, when using screenshots for appstream data you should put
your screenshots in there and reference them via
<a href="" title=""></a>

git clone kde:websites/product-screenshots
<a href="" title=""></a>

Mind the readme.


what to do with repos

whilest doing git repo house cleaning in preparation for KDE moving
repos into phabricator I stumbled upon some stuff that runs under my
name but probably should be moved to launchpad now that its git
feature is being used for kubuntu stuff. please have a look at the
repos and tell me what to do with them

<a href="" title=""></a>
clone of kde-runtime with continues rebase branches of l10n patch.

how to appstream

I couldn't find reasonable documentation on how to best add appstream
data to a kde application, so I threw this together here:

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


Welcome Clive Johnston as Kubuntu Dev

After a marathon grilling I am happy to say that we can welcome Clive
as new member to the Kubuntu Dev team. He's been very active and
dedicated to Kubuntu and we are glad to accept him into the ranks of
elite Kubuntu developers.

Congratulations and welcome!


/: convert all duplicated icons into symlinks

Git commit c994bb7dda00c8bb3713f9d5c75334c3dabdccb8 by Harald Sitter.
Committed on 28/09/2016 at 08:57.
Pushed by sitter into branch 'master'.

convert all duplicated icons into symlinks

I understand on windows symlinks need to be manually flattened, so I
suggest the KDE Windows team makes sure whatever they do for breeze is also
done for oxygen

CCMAIL: <a href="mailto:kde- ... at kde dot org">kde- ... at kde dot org</a>
REVIEW: 129026

T +- -- 128x128/apps/kbugbuster.png
T +- -- 128x128/apps/kmail.png
T +- -- 128x128/apps/office-calendar.png
T +- -- 128x128/apps/plasmagik.png
T +- -- 128x128/apps/utiliti

Snappy sprint reporty musing

Hello sweeties!

Last week Scarlett, Aleix, Matthias and I took to the Snappy sprint in
Heidelberg to discuss how to make it the most useful for KDE. I'd like
to give you an overview of what was discussed along with some
background, so we are all on the same page.

If you already know what Snappy is, you can skip the next paragraph.

Snappy is one of them fancy new bundle packages: AppImage, Flatpak,
Snappy. Out of those three Flatpak and Snappy are based on Linux
Containers/CGroups/Namespace and that whole slew of magic. Simply put
they are like docker.

Review Request 128520: fix incorrect markup usage

Review request for Kubuntu, LibQApt and Aleix Pol Gonzalez.

Repository: libqapt

since kf5 kuit markup needs to be run through xi18n* rather than i18n*

(this really should have a build-time check in ki18n...

The curious case of stuck systemd poweroff


ever since systemd and or sddm started not killing all our session
processes we have had problems of poweroff/reboot getting hung up
waiting for processes to quit.
Recently systemd then started sending them TERM by default, which in
theory should make things behave as before, but more often than not it

The reason for this is meh to debug and altogether somewhat
convoluted. still needed?

<starbuck1> valorie: can someone decide, if <a href="" title=""></a>
is still needed with the new kubuntu webpage "News" section?


Review Request 127821: mark runtime dependency as type runtime

Review request for Baloo.

Repository: kfilemetadata

mark runtime dependency as type runtime

CMakeLists.txt d4b6f7e9a378bbb41aa90252aff6d587ffca35f1

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



Harald Sitter

appstream icons and the hicolor madness we have in KDE software


For the longest time I was annoyed by dolphin having a fishy icon in
discover, which as it turns out is because it uses a default icon name
so it gets a silly icon from a silly theme as appstream has no proper
way of handling this.


That lead me to notice that there is branding identity problem with
how most of our apps handle icons.
Our applications for the most part assume the default icon theme as
their icon, so discover's icon is the current breeze icon, kmail's
icon is the current breeze icon and so on and so forth.

CIs: stricter versioning control

Since the debian version epoch keeps getting messed up every once in a
while in the CI branches I have now rolled out tech to prevent debian
epoch bumps (basically bumping the super overriding version of

Our jobs now record the versions they built in a file called
'last_version', this file is read on subsequent runs and it is made
sure that the epoch version of the new build is not different from the
last one (no increments nor decrements allowed).

If the epoch is found to be different this will result in an unhandled
exception in the log looking a bit like this:

muon 5.5.0

New standalone muon package manager release.

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

* discover, updater and notifier were split. muon only contains the
package manager now
* build dependencies have been reduced accordingly

Kubuntu CI Show & Tell @ Ubuntu Online Summit


It was suggested that I do a show and tell on Kubuntu CI. So I got one
scheduled for Thursday Nov 5 (a week from today).

I'd like to invite everyone to attend. In particular if you are
interested in how our CI works or the concept and motivation behind
continuous packaging.

More info and registration for UOS:
<a href="" title=""></a>


kde test fail in 16.04

<pitti> sitter, Riddell: several KDE tests now fail with "FAIL stderr:
cc1: warning: command line option ‘-std=c++11’ is valid for C++/ObjC++
but not for C"; e. g. kio, plasma-workspace, kcoreaddons
<pitti> these could be quiesced with "allow-stderr", but presumably
this should be fixed more properly?

Would be good if someone could take a look. If you have question I am
sure pitti can help.


CI merge marker checks

It has happened more than once that git threw a merge conflict and in
a hurry someone didn't resolve the conflict properly but simply git
add a conflicting file. This leaves lingering merge markers (>>>>>>
<<<<<<) in the file and potentially goes unnoticed for a long time as
some files are not actually parsed at build time and thus the merge
markers do not cause a build failure.

I now added tech to detect this in the QA stage of all KCI builds [1]
and raise integration errors if markers were found.

kdepim ABI

<a href=";id=c76f2759cc10a2b19138f083da5791858cbd3e1c" title=";id=c76f2759cc10a2b19138f083da5791858cbd3e1c"></a>

it appears to me the only incompatible change is [1] which appears to
be in viewer.h which hasn't been installed in the present soversion
ever, so this is actual private ABI that was changed and doesn't
warrant soversion twiddling of any magnitude.

[1] <a href=";a=commit&amp;h=9d52b22b3652f4dc274114f41c6ea83a31477c01" title=";a=commit&amp;h=9d52b22b3652f4dc274114f41c6ea83a31477c01">;a=commit&amp;h=9d52b22b3652f4dc274114...</a>

unstable CI builds

FYI I switched out our log parser and didn't get done with
reimplementing the ignore rules so expect builds to turn UNSTABLE
because they have missing optional cmake deps. Simply ignore that.
I'll retry all builds next week once I have this bit of tech
reimplemented (needs some code design fiddling first unfortunately).


CI merge order changed again - backports back, order adjusted

Backports branches are back in the game. But the merge order changed somewhat.

So. The reason I originally introduced backports as it turns out was
that we consider them a 'release' branches. i.e. we release from this
branch (albeit after distro release). As such it plays into CI efforts
of that distro series (e.g.

pause CI before mass pushing stuff & mark irrelevant changes NOCI

General reminder:

Use the tech to avoid transitional CI fail. Specifically pause CI when
you push lots of plunder via
<a href="" title=""></a> and use NOCI in your
commit messages when you want to prevent an immediate build.

# Pause
<a href="" title=""></a>
This job halts all build integration while it is running. Builds, and
only builds. You still get semi-instant feedback on broken merges and
so forth.

CI merger order change - backports remove

I have just removed kubuntu_*_backports branches from the CI
auto-merge order. In my original introduction of it I failed to
explain why we want it and I am now rather under the impression that
we do not want it as it basically regresses wily branches to backport
conditions ever so often while not actually adding any value on top of


CI dep resolvers

Quick brain dump.

We are currently using pbuilder's dependency satisfiers, but they are
rather error prone for CI purposes.