Venue

enioka
13 rue du Mail - 75002 Paris - France
To get to the venue:
- Pass the large wooden door on the street
- First door to the left to get in the stair case
- Get to the third floor, ring at the door marked enioka
- Someone (likely ervin) will open

Agenda

Add topics we should discuss:

Progress report on the topics from 2024 https://invent.kde.org/groups/pim/-/milestones
Switching to the sqlite backend by default
How to get more information to debug weird cases in the field
Enabing KUserFeedback to get survey data from users
Design a survey or surveys to know our (former) users better
Building criteria and buckets on how to handle features removal
Look at the channel to contact us for bugs or other requests, see what makes sense for us (we have several and they all feel like /dev/null from the outside)
Find a way to tame the "random features popping up" effect
Status report on KMime frameworkification
Documenting the port of agent/resources settings GUIs to QML
KDE PIM on Flatpak status quo and further plans
Online Accounts System

Notes

Kolab resource retirement

Status: The resource has already been dropped by some distributions because it does not longer build with current libraries (e.g. Boost).
The proposed solution of notifying existing users to manual switch to IMAP/DAV is implement since a sprint or two ago, but wasn't merged yet.
Migration isn't going to get implemented, and hard, so waiting for that makes things actually worse.
Plan: Add a dummy resource that shows a notification with hints for migration.
See also https://community.kde.org/Sprints/PIM/2024#Kolab_resource_retirement

KMime frameworkification

Status: https://invent.kde.org/pim/pim-technical-roadmap/-/issues/49
Plan: Get it into KF 6.23 (release February 2026) and switch KDE PIM to it with Gear 26.04

Online Accounts System

Nico demo'ed the current state of this (https://invent.kde.org/system/konlineaccounts)
Some remaining work: https://invent.kde.org/system/konlineaccounts
What to do about the old online accounts code/uses? Goging to be removed.
Nico will finish this, will ideally be part of 26.04
Application side can be integrated already, as this has to gracefully fall back when the D-Bus daemon isn't available anyway.

Use KUserFeedback

KUserFeedback is already used by KMail, ...
include types of used resource and plugins
- filtering out unknown resources/plugins would make that safe
settings are more tricky, we'd probably need a proper pull infrastructure for this first in KUserFeedback
Survey: what would we want to ask? and what would we want to actually change based on the result?
- Would make most sense when the answers actually have consequences, ie. what would we change/do with the results

KDE PIM on Flatpak

Kontact one exist, Merkuro in review, with duplicated Akonadi
Can we have a shared Akonadi Flatpak? currently not, we would want a Flatpak Akonadi runtime that also contains the server processes (that's non-trivial to implement though, with bigger implications for sandboxing and permissions etc, as we need also file system and socket access, not just D-Bus)
Mulitple Akonadi instances in different Flatpaks can't share data and make host integration difficult
- probably we should set this up with separate Akonadi instance identifiers
We'd want a KCalendarCore plugin that can use a Flatpak'ed Akonadi instead of a host one
Check how GNOME does this, they seem to have a similar situation in PIM nowadays
We need a KCalendarCore::Calendar interface over D-Bus to bridge between host and flatpak instances either way, as the raw Akonadi protocol isn't stable / ABI compatible
we could explore a custom setup for a service Flatpak containment, as a first step/prototype, instead of doing this upstream first (lacking another usecase there fore now), based on a shared Akonadi runtime containing the libraries and binaries

Switch Akonadi to the sqlite backend by default

Default on new Neon installation since 2023
Kevin is using it for a while, Carl on one new installation.
Fails on super large queries, but we want to get rid of those anyway. Carl fixed all known ones already.
For now it's only about new installations, migrating comes later, to in the very long run drop support for the other backends.
Can we make the migrator to have a --test-run or --dry-run option that just writes a new sqlite db to test the process
Drop support for migrating to Postgres already, as that requires a fixed Postgres version

How to get more information to debug weird field issues

Attach relevant metadata to sentry reports, assert on more internal invariants to actually get Sentry reports
Can we get the resource identifier set as process name in journald logs
More structured log output containing job ids/transaction ids
should go down in the same way to kimap, kdav, kio, etc
All of Akonadi gets attributed in systemd to the process that first starts it, which is usually the reminder daemon. Probably solvable with a systemd unit file for akonadi_control (fixed meanwhile by Nico)

DAV collection loss isusue

the KIO error handling fix helped, but doesn't fix this entirely
Kévin found another issue in KDAV where error propagation fails due to some "clever" workaround for some broken server it seems.

QML resource configuration UI

plugin porting done partially during a GSoC
how does the current account wizard relate to the new online account system
online account system would need a "generic email" acount type, which gets the auto-discovery logic from the account wizard

Address auto completion improvements

default blacklist pattern @noreply|noreply@
see akonadi-search krunner for more patterns

Sprint Schedule

December 13/14, 2025
It's encouraged to show up to the venue on the Friday 12 December afternoon
Kevin will be already there to greet the early birds, work can be started when enough people show up

Travel Reimbursement

File a request on https://reimbursements.kde.org/events/222
The proposed hotel is around 160€ per night.

Attendance

Kevin Ottens
Volker Krause
Ingo Klöcker
Carl Schwan
Nicolas Fella
Florian Richer
Albert Astals Cid
Tobias Fella
Laurent Montel

Hotel

Plenty of options in the area, this is Paris after all... prices can quickly go up though, the proposed hotel is one of the cheaper ones with proper availability
Proposed hotel is "Hôtel Chopin": https://hotelchopin-paris-opera.com/
It's ten minutes by foot from the venue
Rooms start at 160€ per night, they can accommodate two persons
There's also a triple room option at 190€ which can accommodate three persons
They still seem to have availability at time of writing but it's already limited

Fallback hotel options

In case the Chopin is full when you try to book, here are other options:

Hotel Ronceray, a good option as well, lots of rooms, floating prices though so can turn out cheaper or more expensive than Chopin depending when you book: https://hotelroncerayopera.com/en/
Hotel Korner Opera, a bit closer to the venue, less rooms (around same size as Chopin), floating prices warning applies: https://en.hotels-korner.com/hotels/hotel-korner-opera

Otherwise you can try to find something in the area by yourself, for instance to keep the price down. Just a word of caution though, if it's under 140€ per night in that area during that time of year, double check it's not a shady place. I thought I managed to get a very good deal once and I regretted it... it was dirty and broken (and I can go on, I can share details another time).

Arrival and departure times

Kevin will already be in Paris on Monday, stays until the next Monday
Volker: arrival Friday afternoon (~5pm at Gare de l'Est), return Sunday afternoon (~4pm from Gare du Nord).
Ingo: Friday noon until Sunday late afternoon
Carl, Tobias and Nico: arrival Thursday, returning on Monday
Laurent: Arrival Friday afternoon, return Sunday in afternoon
Albert: Arrival Friday afternoon, return Monday evening

Sprints/PIM/2025