Stephan Assmus wrote: > > In case it is really not working like on BeOS, then it is a bug. The > registrar was developed much earlier before the app_server matured, > so we > might have well overlooked something. I figured it was a bug, but you never know, so I asked. > I would guess that B_SOME_APP_ACTIVATED is for other apps, while > B_APP_ACTIVATED is for your own app. Only the latter is supposed to > trigger > the AppActivated hook to be called. Are you sure that any app is > supposed > to received B_SOME_APP_ACTIVATED? Or does it specifically need to ask > the > roster it wants to receive this notification? I'm not exactly sure what Be Inc. *meant* to do, but right now B_SOME_APP_ACTIVATED is sent where B_APP_ACTIVATED should be, i.e. when the application in question receives or loses the focus. To follow the documentation, we should do the same but actually use B_APP_ACTIVATED. B_SOME_APP_ACTIVATED is supposed to be used for the app watching stuff. Looking at our Registrar code, it is correct, but we need to add the link between the app_server and registrar for the activation. Right now it seems only the app_server is involved, so the activation code in the Registrar goes unused. In fact we need a two-way link so that I can fix the above Launch bug (activate the app window on launch if it is already running.) > Ok, what about the shutdown sequence thing? Should someone else fix > the > input_server shutting down, or did you mean to do that yet? I don't > mean to > be pushy or anything, just asking... :-) Don't worry I'll fix that too. For now the previous code I submitted won't hurt anything. I just keep finding and investigating bugs and if they seem simple I prefer to make a quick fix. I thought this activation thing would be easy, but it is trickier than I thought, especially with the buggy BeOS behavior too. It is hard to figure out what is "right" when the OS we are recreating does it wrong, partially at least ;) Regards, Ryan