On 24/07/2020 20:38, Patrick Ohly wrote:
[Most of the text below was written in December 2019, but than
unintentionally sent to an internal mailing list - no surprise that I
never got any response!]
Over the Christmas holidays I worked on building a new SyncEvolution release. My
current goal is to build for Ubuntu Bionic (most
recent LTS) and support those binaries on all more recent Debian and
If possible, I'd like to drop unused features if they require extra
effort. This mostly depends if someone still needs them. Let me list
some features that I'd like to remove. If you still need them, please
* At the top of that list is ActiveSync support. activesyncd no longer
builds on Debian Stretch because it depends on libgnome-keyring, which
was removed. It probably can be ported to libsecret, but that's
I no longer need ActiveSync myself (my employer banned it a while ago
and, more recently, I have retired anyway). I feel it is a shame to let
it die and could probably be persuaded to do some work on it if someone
actually wanted it and could provide me with access to systems to
develop and test. But if no one wants to use it, there is no point.
* x86 (i.e. 32 bit) binaries - it doubles the testing effort.
* RPMs - they never had proper dependencies and I am not sure whether
they ever worked at all.
* Akonadi support and KDE in general.
I first encountered problem with Akonadi in Debian Stretch and reported
it here with a stand-alone reproducer:
But as pointed out in that issue, the API that SyncEvolution uses is no
longer supported and thus SyncEvolution would have to be ported to the
current API, whatever that is - I haven't investigated that.
I gave up on the KDE PIM tools (and associated data) some time ago so I
am no longer using this.
* Port to Python 3 and stop supporting Python 2.
Regarding the source code, I'd like merge all pending patches. This
obviously includes all the changes that are required to build on more
recent Linux distros, but also the C++ modernization that I started a
The result will be more than just a simple bug fix release, but also not
something that has any new user-visible features. I'm not entirely happy
with that, but I also don't want to be stuck completely in pure
I got testing on the newer Linux distros working with the updated code
base already beginning of this year, but then got stuck because of a
regression and lack of time to dig into that. Since then, the apt repo
keys expired and I haven't renewed them because the binaries probably
wouldn't work anyway.
I suppose users would like to see binaries again, primarily because
SyncEvolution fell out of Debian/Ubuntu?
Now that I am retired, I am not sure what my usage will be. My (new)
master for contacts and calendar is in Owncloud/Nextcloud (mainly using
Thunderbird as UI), however I haven't decided how I will keep various
other devices synced (particularly Sailfish, if I continue to use that).
I would find it convenient to have Debian binaries although I don't know
if I will need them long term.
I am disappointed, having worked on PIM sync for many years, that the
world seems to be willing to settle for very limited and mostly locked
up services from Microsoft, Apple or Android.