https://bugs.kde.org/show_bug.cgi?id=443984
--- Comment #2 from Nate Graham <n...@kde.org> --- ...And that calls KAuthorized::authorizeControlModules(), which only checks for whether there are authorization restrictions in kdeglobals under "KDE Control Module Restrictions". So according to this function, it is perfectly valid to report that a module is authorized even if it is not installed. Options for fixing this 1. Change the definition of "authorized" to include "installed" in KAuthorized::authorizeControlModules() 2. Change the definition of "authorized" to include "installed" in KCMShell::authorize() 3. Create a new function KCMShell::authorizedAndInstalled() that uses KCMShell::authorize() but also checks installation status, and port all uses of KCMShell::authorize() to it 4. Add an additional check for installation status in every QML applet that uses KCMShell::authorize() Option 1 would be easiest and involve the least new code and porting churn, but would entail changing the what the function returns in ways that callers might theoretically not want. Though it's hard for me to even imagine a use case for calling KAuthorized::authorizeControlModules() on KCMs that aren't installed and actually wanting it to return true. I suspect if I submit a merge request for this, Frameworks people will complain about it. But maybe not. Option 3 is probably the most correct, but once we did it, we'd port everything that uses KCMShell::authorize() to KCMShell::authorizedAndInstalled(), begging the question of why anyone would want to use KCMShell::authorize() anymore given that it doesn't also factor in installation status. Albert, what do you think? -- You are receiving this mail because: You are watching all bug changes.
