https://bugs.kde.org/show_bug.cgi?id=488870
--- Comment #19 from rallyemax <kdeb...@syrinx2112.addy.io> --- (In reply to Martin from comment #18) > I am hip-firing here while blindfolded without reading much of the codebase, > but if you change it to Qt::Key_CapsLock instead of trying Enter, does it > work? > > If yes, > https://invent.kde.org/libraries/plasma-wayland-protocols/-/merge_requests/ > 82 is relevant. > > If not, sorry, was just digging through both that PR and this issue by > happenstance. YES! Changing to Qt.Key_CapsLock eliminates the error. Changing to any of the keys proposed for addition in that MR causes the error to reappear. What confuses me is why merely lifting the element two levels in the hierarchy also prevents the error from appearing. But that feels like a red herring now, and likely due to my limited understanding of QML. Nice catch! I was trying to chase it down and got as far KModifierKeyInfo in the KGuiAddons package. I got a bit stumped figuring out how the m_modifierStates hash gets populated. It looks like it gets populated from the keystate Wayland protocol you linked. I just can't keep up with these deep hierarchies of abstractions anymore. So it looks like this set of changes (https://invent.kde.org/plasma/plasma-workspace/-/commit/72f1018f0c989ed2b316dff487c3b517b6b2e89d) changed the key lookup mechanism from something that worked to the incomplete keystate protocol that no longer does? -- You are receiving this mail because: You are watching all bug changes.
