Copyright © 2019, 2022 Collabora, Ltd. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice (including the next paragraph) shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. Starting with version 2 of the protocol, the client is required to wait for the 'bound_ok' or 'bound_fail' events in order to proceed further. In case the client gets a 'bound_fail' event then it should consider that there's another client already bound to the agl_shell protocol. A client that receives a 'bound_ok' event should consider that there's no other client already bound to the interface and can proceed further. If the client uses an older version of the protocol it will receive automatically an error and the compositor will terminate the connection, if there's another client already bound the interface. If the client receives the 'bound_fail' event and attempts to use the interface further it will receive an error and the compositor will terminate the connection. After the 'bound_fail' event was received the client should call the destructor, which has been added with version 2 of the protocol. The client is free to try at a later point in time to see if it will receive the 'bound_ok' event, but there's no explicit way of finding out when that event will delivered. It is assumed that it can infer that information through other means/other channels. Tell the server that this client is ready to be shown. The server will delay presentation during start-up until all shell clients are ready to be shown, and will display a black screen instead. This gives the client an opportunity to set up and configure several surfaces into a coherent interface. The client that binds to this interface must send this request, otherwise they may stall the compositor unnecessarily. If this request is called after the compositor has already finished start-up, no operation is performed. Set the surface to act as the background of an output. After this request, the server will immediately send a configure event with the dimensions the client should use to cover the entire output. The surface must have a "desktop" surface role, as supported by libweston-desktop. Only a single surface may be the background for any output. If a background surface already exists, a protocol error is raised. Set the surface to act as a panel of an output. The 'edge' argument says what edge of the output the surface will be anchored to. After this request, the server will send a configure event with the corresponding width/height that the client should use, and 0 for the other dimension. E.g. if the edge is 'top', the width will be the output's width, and the height will be 0. The surface must have a "desktop" surface role, as supported by libweston-desktop. The compositor will take the panel's window geometry into account when positioning other windows, so the panels are not covered. XXX: What happens if e.g. both top and left are used at the same time? Who gets to have the corner? Only a single surface may be the panel for an output's edge. If a surface already exists on an edge, a protocol error is raised. Ask the compositor to make a toplevel to become the current/focused window for window management purposes. See xdg_toplevel.set_app_id from the xdg-shell protocol for a description of app_id. If multiple toplevels have the same app_id, the result is unspecified. XXX: Do we need feedback to say it didn't work? (e.g. client does not exist) Ask the compositor to hide the toplevel window for window management purposes. Depending on the window role, this request will either display the previously active window (or the background in case there's no previously active surface) or temporarily (or until a 'activate_app' is called upon) hide the surface. All the surfaces are identifiable by using the app_id, and no actions are taken in case the app_id is not/was not present. See xdg_toplevel.set_app_id from the xdg-shell protocol for a description of app_id. Informs the client that it was able to bind the agl_shell interface succesfully. Clients are required to wait for this event before continuing further. Informs the client that binding to the agl_shell interface was unsuccesfull. Clients are required to wait for this event for continuing further. Notifies application(s) when other application have suffered state modifications. The agl-shell protocol makes the windows by default maximized to the output area, taking into consideration the panel sizes. If certain client would want be behave like a pop-up type, always sticky window, when switching between applications, it can use this request to do so. If the application, specified with the app_id string, is not currently running, it will be added it to a pending list. If, at a later point in time, that application does come up it will apply the floating state to it. Note that once that happens, and a later point in time that application is stopped, these changes will not apply anymore. To keep a permanent state see also set_app_property_mode(). Another use-case is for applications that want to be started from the beginning as floating, so they the client must call this request before doing the initial wl_surface.commit(). Application can also transit from float to maximized and vice-versa using this request or the set_app_unfloat to go back to an initial maximized state. See xdg_toplevel.set_app_id from the xdg-shell protocol for a description of app_id. This undoes the effect of the 'set_app_float' request in case the client wants to go from floating back to the maximized state. If there's no app_id present this request will obviously do nothing. See xdg_toplevel.set_app_id from the xdg-shell protocol for a description of app_id. See xdg_toplevel.set_app_id from the xdg-shell protocol for a description of app_id. Use this request to inform the compositor to maintain a pending state for an app_id being set with set_app_float/set_app_remote request. Any subsequent application matching that app_id would survive a potential application destruction. Note that this request will take effect globally on all applications. To turn it on, or off, use the 'permanent' argument. Initially, the compositor will have this option set to off. Note that it doesn't matter the order of this request with respect to set_app_property() request, as the changes will only take effect when the application itself does the commit with an app_id set, therefore the only requirement is to call this request before the app_id client does its first commit. Request changes the application from the original mode (whatever that might be) to a split, tiled orientation mode defined in the orientation split_orientation enum. This request only handles a single level of orientation for practical reasons. If there's no app_id with the supplied name this request does nothing. A none orientation type would make the window go back to the original maximized mode. This implies that either the background surface will displayed or in case there are more than one applications being shown, the other application will be returned to the original, maximized state. The tiled orientation could be applied independently of each other, such that a client can transition from one tiled orientation to another. Any other window already present would switch places with the currently one being changed. In case there's already a client that's already active at that time, it will be attributed automatically the opposite tiled orientation, such that two concurrent applications can be displayed at the same time. The orientation tiles can not be combined, and only state at a time can be active, which means we can't have a top left orientation, or any combination like that. See xdg_toplevel.set_app_id from the xdg-shell protocol for a description of app_id.