On Fri, Feb 28, 2014 at 9:48 AM, Alberto Mardegan <alberto.marde...@canonical.com> wrote: > On 02/28/2014 10:30 AM, Thomas Voß wrote: >> Nope, the trusted session approach is different in that it will be not >> available to ordinary clients. We do not allow the client to be able >> to interfere with the stacking order and instead enable a trusted >> helper to: >> >> (1.) Create a trusted session and add participating apps and surfaces to >> it. >> (2.) Start the trusted session. At that point, Mir & Unity8 take >> over and carry out a well-defined set of operations, including >> on-screen transitions. >> (3.) Be notified when the trusted session ends. > > I hope to be proved wrong, but I've the feeling that this can be racy: > we need to make sure that the trusted helper manages to start the > trusted session before the participating apps have shown their windows, > or you won't be able to see the correct transitions.
Why not? The designs explicitly take into account the case when an app is already running. Once the session is started, Unity8 takes over and ensures seamless transitions and tight integration with the lifecycle. That is, calling start on the session completely hands over control to the shell, which is the synchronization point then. > I think that this needs some cooperation from the application, meaning > that it needs to be aware of the fact that it's taking part in a trusted > session. > I disagree. Applications taking part in trusted sessions will be able to define custom exec lines for specific use-cases (say content-picking), but will not receive any further information about the overall system state. >> All of this is entirely transparent to an application and there is no >> need to expose this to the toolkit at this point. > > I'm selfishly thinking only about the trusted helper now: I've currently > code for window reparenting using QWindow::setTransientParent() which > works on X11 (because the xcb QPA plugin handles that for me), and it > would be nice if I could reuse the same code. I don't think we will be able to leverage the existing code here as the underlying concepts differ significantly. Sorry for that :) Cheers, Thomas > If the trusted session > were handled via the Mir QPA, I would avoid a dependency on Mir and a > few #ifdef's. It's not a big issue, but a nice to have. :-) > > Ciao, > Alberto > > > -- > Mailing list: https://launchpad.net/~ubuntu-phone > Post to : ubuntu-phone@lists.launchpad.net > Unsubscribe : https://launchpad.net/~ubuntu-phone > More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
