Details
- Reviewers
- None
- Group Reviewers
Plasma - Commits
- R119:f2e579f9b366: Integrate the normal keyboard layout switcher
I can switch between catalan and american keyboard, if ibus
is running I can do the same but through ibus.
Diff Detail
- Repository
- R119 Plasma Desktop
- Branch
- master
- Lint
No Linters Available - Unit
No Unit Test Coverage - Build Status
Buildable 17455 Build 17473: arc lint + arc unit
layout-templates/org.kde.plasma.desktop.defaultPanel/contents/layout.js | ||
---|---|---|
38 | I'd say so. We can do it when we deprecate the SNI though, unless you really want it now. I've never done it so I'll have to research how it's done. |
It's a small part of the work Geon Park did in his GSoC project to refactor kimpanel and our input sources management in general, but in a more merge-friendly format.
The former sadly never got any review attention because the patch was too big, so maybe this is the right way to make progress.
For context though, in the GSoC the following work was done:
- The keyboard layout KCM was rewritten in Qt Quick
- It's based on a model that can contain both IMEs and keyboard layouts
- The model can dynamically react to IME daemons coming and going and doing a sensible fallback
- The implementation is shared between the KCM, kimpanel (which, like here, replaces the old SNI) and the kded module that on X11 provides the D-Bus API to manage layouts
- That means global shortcuts like Next/Prev Keyboard Layout we have work over both IMEs and layouts
- That means the OSD and the lock screen show the right thing and became IME-aware
And a couple of other things. What wasn't done was touch KWin/Wayland (which in a Plasma/Wayland session provides the same D-Bus API kded provides on X11).
So it's important to remember that this is the goal. With this patch Plasma would get better, but it doesn't address the whole problem (System Settings being useless for IMEs, the IME daemon making System Settings useless when running, two incompatible systems fighting each other, expert knowledge required to set things up in the first place, etc).
I'm very excited to see this because it indicates more interest growing in the problem domain, so maybe we can review and merge that other body of work down the line after all.
Can that work be resumed/used?
See D14796. It seems like the original author is MIA. I rebased the patch, seems to work but adopting it will set us back as the ibus backend was never implemented.
I'd say so. We can do it when we deprecate the SNI though, unless you really want it now. I've never done it so I'll have to research how it's done.
I don't want it to get forgotten and us to do work in a rush when the release comes along.
As for the other IM work, that was the state when I originally wanted to port this layout item (T5331) 2.5 years ago. Doing everything all at once isn't an approach that seems to be working.
Having the kimpanel on by default and dropping the SNI from the kded/kwin seems like it's a step that would be needed anyway, so this is work towards that end goal even if this specific plasmoid does end up getting replaced.
From my POV, ship it.