hein added a comment.
> I'm possibly wrong, but instead of the new set couldn't you > if (appDataCache.contains(window) && appDataCache[window].icon.isValid()) I don't understand on multiple levels :) There's no QIcon::isValid, and even if there was, the icon being in the cache and being valid could be the result of the the preferred codepath (from .desktop file) in which case we don't want to emit DecorationRole based on a NET::WMIcon prop change. Basically this patch is about "react to NET::WMIcon for windows we actually show it for, but only those". REPOSITORY R120 Plasma Workspace REVISION DETAIL https://phabricator.kde.org/D7092 To: hein, #plasma, davidedmundson Cc: plasma-devel, ZrenBot, progwolff, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas, apol, mart, lukas
