In terms of the UI, we're trying to move away from tabs and towards multi-page KCMs where clicking on a button takes you to another page. For examples, see the Notifications and Style KCMs.

Thanks for the comments.

In T12831#223710, @alexde wrote:

For inspiration: I think Piper handles the key mapping pretty:

Hey guys I'm back.

Phase 1

In D13997#289384, @romangg wrote:

If we go with this split the stages should be:

1. Append an x to the name of the old Touchpad KCM, make the KCM show only on X and add the new Wayland KCM without the X code as kcm_touchpad.
2. Remove all Wayland related code from the old KCM.
3. Rewrite the new Wayland KCM as KConfigModule.
4. Your stage 4.

Stage 3
kcm_touchpad (Wayland) and kcm_touchpadx (X11) are available by this commit. Tested and working fine. To test it, you should manually remove the old kcm_touchpad files since the names are clashing imho. This problem will be handled by packagers after the patch is pushed to master. I'm implementing the backend now (kcm.cpp, kcm.h).

Stage 2 :
Minimum changes are done since there will be another commit for kcm_touchpadx cleaning. The kcm_touchpad is changed as kcm_touchpadx and it is moved to new folder "touchpadx". Since "touchpad" folder is changed, you can review the diff by checking it. Tested on X and everything is working fine. In Wayland, the KCM is not shown at all.

Since you can review the old diff, I'm moving to stage 2. If you have problem with the Stage 1's diff, please tell me.

Comment correction.

"value" part of last "Slider" patch is reverted since that kind of changes will be handled with another patch.

In D13672#281774, @ngraham wrote:

BUG: 395681 can't be a link, I don't think. It's gotta just be plain old dumb text. :)

This patch actually solves the problem specified on bug title, "libinput pointer kcm needs 11 ticks on the speed slider so that the middle one can correspond to 0.0". Still you think that it doesn't solve the bug ?

Typo fix.

In D13573#280026, @ngraham wrote:

Looks good to me too.

In D13573#279852, @furkantokac wrote:

• Please review it. After your review, I'll also enable the Acceleration Profile with this patch by some editing.

That would be unrelated to this code cleanup. Please do that in another patch.

V0.2

• Converted everything to qqc2.
• SizeHint is corrected.
• ToolTip.qml is deleted since qqc2 has built-in support for ToolTip as we want.
• Tested. Overviewed all code.
• Please review it. After your review, I'll also enable the Acceleration Profile with this patch by some editing.

@davidedmundson @mart
Thanks for the feedback. New patch is coming.

It would be very nice! I have things to do that will take 1-2 month. If nobody interested till then, please ping to me. I can work on it.

V0.6
Wayland version (main.qml) is tested (X was already tested). Corrections are done. Functionality is same. All the possibilities are tested. Works fine. Ready to push.

V0.5
import org.kde.plasma.core is removed for main.qml.

V0.4
QtQuickControls version is updated to 2.0. Some code cleaning.

In T8985#146933, @abetts wrote:

Let me know if I can help with anything

In D13372#277078, @mart wrote:

here a patch against this which ports it to qqc2, using the completely different radiobutton mechanics qqc2 uses, fixing separators, removing plasmacore import

qqc2.diff3 KBDownload

with qqc2-desktop-style master, the spacing issues should be solved

In D13372#277040, @mart wrote:

for me, using qqc1 in a new kcm, is an absolute blocker, no matter what the problem is with qqc2, problems must be fixed, not worked around. ever.

V0.4
Stupid mistake that I have done is fixed. Thank you romangg.

V0.3
sizeHint and minimumSizeHint are added.

If everyone is happy and the patch is tested, I'll push the code after a last test for Wayland and Xorg.

V0.2
After a discussion, decided to go back to V0.2 because of some UI issues. Bug reports will be filed about the issues. For now, we'll use qqc1.

Just realized that paddings look bad in QQC2. Any recommendation ?

V0.3
QtQuickControls version is updated to 2.0. Some code cleaning.

In D13372#275788, @safaalfulaij wrote:

Small question, why we're not using QQC2-based controls?