Planner project status

[Mart Raudsepp <leio gentoo org>]

Hello!

I realized I should probably give a status update on planner on here.

While I have your attention to start with — to go along with GNOME
infrastructure changes, this and the user mailing list may get
decommissioned in the not too distant future, in favor
of https://discourse.gnome.org/. Once we are actually starting this
migration, there will be a posting detailing that. No mail subscription
transferring will happen, so you will need to set up things on
Discourse for yourself at that point to see planner-related postings.

Now onto the main topic.


Half a year ago I became the new maintainer of planner.
Before that and since then, I and some other contributors have been
working on various technical debt, most of which has now been
integrated into the git master branch. Some noteworthy mentions:


* All issues and feature requests have been migrated from bugzilla to
GNOME gitlab: https://gitlab.gnome.org/World/planner

* The UI code has been ported to use GTK3 instead of GTK2. As pre-
requisites to that, patches to migrate from GConf to GSettings (dconf)
have been included, and the UI descriptions have been migrated from
libglade to modern GtkBuilder files.

** Additionally, this meant having to move from old GDK drawing API to
cairo for drawing, which means that various drawing in the gantt
display should look much better visually (anti-aliasing, etc).

** libgnomecanvas is still used, but as a bundled GTK3 supporting
version, borrowed from evolution mail client with further work on top
in order to bring back a few small things that planner uses but
evolution doesn't and thus had been rightfully dropped during their own
porting to GTK3 (long ago).

* Initial work on porting to the meson build system was picked up and
brought to completion. Autotools is not used anymore. Nor is intltool.

* Patches to fix building against newer evolution-data-server and GDA
have been integrated.


Some things have been (intentionally) lost as well:

* Public libplanner library
* And with that also support for external plugins, including python

That's because we need to modernize things and the libplanner API is
therefore going to be unstable for the mid-term future, thus it makes
sense to not commit to a public stable API for it after a decade of no
releases for the time being.
On the other hand, and perhaps more importantly, I'm not sure we even
need such a public API at all at this time or foreseeable future, if
ever. If there are actual use cases for this, we can revisit it later,
not immediately to hinder the refactoring. I believe many cases could
be handled by integrating it into planner proper or as a plugin that
lives in planner source tree and uses the private API.

If you used some external plugin that is necessary, please let me know
so we can see how to retain the functionality it provided.



I hope we can make some sort of an alpha releases with the GTK3 port
soon - the main known problematic bit is some bad date picker popup
behaviour within the task lists, on Wayland in particular (yes, it
works natively on Wayland now with GTK3).

I would welcome extensive testing of the current git code to find out
any other issues there may be.
I must admit that I'm currently not really a planner user and only
briefly dabbled with trying to use it for some project planning in the
distant past. I've been working on it mostly out of interest in the
work itself. Therefore it would be great to have more knowledge within
the active project community on the actual planning aspects, WPS and
all that stuff as well and I invite you to join me in the Planner IRC
channel. Information about that is available on
https://wiki.gnome.org/GettingInTouch/IRC - we use the channel named
#planner
Things tend to go slow for me when I don't have active company, so just
chatting about planner-related things there might speed up progress as
well.  So would actively helping with code, help files, and so on ;)


As for future short-term plans for the project, this is my current
rough list in my mind:

* Fixing any major GTK3 regressions, like the aforementioned date
picking issue
* Changing the main branch from `master` to `main`
* Releasing an alpha version to get more users to test it
* Work on restoring Windows and Mac build/installer support
* Release a beta version with installers/packages for Windows/Mac
included
* Releasing a new "stable" version
* Working on modernizing lots of things further (currently done work
has been really the bare minimum to get it to even compile against
GTK3)
* Start looking at UI redesign and new features

And again - thoughts and brainstorming partners would be very good
going forward.


Best,
Mart Raudsepp