wed, 15th 3pm cet there will be a short meeting about porting plasma
to layershell

David E
* I have started a replacement for the datetime dataengine with simple
QML bindings - it's going to be amazing
* digital clock gained a lot of hacks that I'm trying to avoid this time
* I should have that ready in a few days
* has same features as timezones and alignment between instances
* a topic I want to dicsuss in teh meeting is this plasma_use_qt_scaling env var
[Nico] The big problem was that it breaks mixed-dpi setups quite badly
yeah, the noteworthy things are:
    we need to have a consistent story in KWindowSystem
[notmart] should we get rid of the env var now to "force" more fixings there?
[ahiemstra/nate]: to me that makes sense break things early so people
run into the issues and they get fixed
[nico] Do we have an idea of what exactly needs fixing in kwindowsystem?
[d_ed] Not explicit, right now the state is we have a few paths that
happen to be used sometimes converting
and some paths that might not be used not converting things
I'm hoping in terms of public api we can end up at :
    KwindowSystem - > logical
    KX11STuff -> device
[nico] There isn't much public API left in KWindowSystem anyway, Most
moved to KX11Extras
[d_ed] ok, I'll see if we can take a look and make some more concrete tasks

* One more superfast topic
* I was thinking of introducing a toplevel folder called "libs"

* We haz Qt6 in Plasma now \o/
* Things don't seem too bad, I've been running it all week
* Some related thoughts:
* There's a couple of Plasma-ish repos not ported yet, e.g. various
bigscreen stuff
* Related to that I sent an email about portentially cleaning up some
repos under plasma/:
* We now have a 'qt6' keyword in bugzilla to tag Qt6-related stuff
I filled that with a couple of crashes already:
* Should we be more systematic about writing down known issues?
[nate] we'll need to be systematic eventually so IMO we might as well
start sooner rather than later
[nico] I was mostly talking about specifically
[d_ed] It might be easier to start with a clean slate and look what
here is actually relevant

    Bugfixing some things creeping up from Plasma 5.27
        memfd screencasting will always work for the first time. A
number of things were fixed as some users pointed out which legacy
clients (slack!) to test with.
        There were some regressions on libkscreendpms on some corner cases.
    Discover port to Qt/Plasma 6. It's kind of there, needs some
dependencies to catch up.
    Worked with Arjen on KPipeWire API that sends ffmpeg-encoded
frames to be used on his new krdp test app.

* I've mainly been working on rdp remote desktop stuff
* now looking into some optimisation stuff so there isn't as much lag
as well as making sure things like resolution are actually properly
done (rather than hardcoded which is the current state)

* removed and ported away from the deprecated "AppletScript" api in
libplasma, merged patches about that
* new repo plasma5support in workspace to split out all the deprecated
dataengine stuff
* ported usages in p-w, p-d and kdeplasma-addons. some more usages may
still lingers in applets in other repos
* preliminar port of plasma-sdk to kf6, plasmoidsviewer works, still
to fix qml of otehr apps here and there
* I still have some potentially breaking changes in the pipeline that
i hope to get them as soon as possible
* changing who the "plasmoid" object is as seen from qml
* split svg in own framework (whether before or after removing the own
scaliong, not sure)
[fusionfuture] If it's time to remove PLASMA_USE_QT_SCALING and
fractional scaling on Wayland is a thing, perhaps
can be considered again
[notmart] probably to try again after one of the two scalings is gone
to make things less headache

* I'd like to have a discussion about porting plasma to layer shell.
Either some time this week or later
Currently not high prio because of qt6 porting
[d_ed] Lets do so in a week, we can start introducing some classes
which have the right plasma-facing API
decision: wed, 15th 3pm cet

* I'm working on killing the kded part of kscreen. The plan is to have
it be X11-exclusive, and have KWin manage the output configs on
Wayland going forward

Marco Martin

Reply via email to