https://bugs.kde.org/show_bug.cgi?id=403172
--- Comment #2 from LinG <lingtj...@hotmail.com> --- Interesting, I assumed that they would use the same call to send the client to a different virtual desktop (Workspace::sendClientToDesktop()) but then only add an additional call in case 5c where you call a method to change the current virtual desktop. I can only give you my point of view as an API consumer, but it seems inconsistent to me that connecting to the same signal sometimes does and sometimes doesn't allow changing the geometry of the client. Especially since changing any of the other properties such as opacity is picked up by KWin without a problem. As far as I'm aware, the geometry is currently the only client property that can't be changed when the desktopPresenceChanged(KWin::Client *client, int desktop) signal is invoked in cases 5a/b. In my personal opinion, it would be better if the API did allow changing the geometry in case 5a/b since they are all handled by the same signal from an API consumers point of perspective. -- You are receiving this mail because: You are watching all bug changes.
