Struct wayland_client::backend::ReadEventsGuard

pub struct ReadEventsGuard { /* private fields */ }

Guard for synchronizing event reading across multiple threads

If multiple threads need to read events from the Wayland socket concurrently, it is necessary to synchronize their access. Failing to do so may cause some of the threads to not be notified of new events, and sleep much longer than appropriate.

To correctly synchronize access, this type should be used. The guard is created using the try_new() method. And the event reading is triggered by consuming the guard using the read() method.

If you plan to poll the Wayland socket for readiness, the file descriptor can be retrieved via the connection_fd method. Note that for the synchronization to correctly occur, you must always create the ReadEventsGuard before polling the socket.

Implementations

impl ReadEventsGuard

pub fn try_new(
    backend: Arc<Mutex<Backend>>
) -> Result<ReadEventsGuard, WaylandError>

Create a new reading guard

This call will not block, but event callbacks may be invoked in the process of preparing the guard.

pub fn connection_fd(&self) -> i32

Access the Wayland socket FD for polling

pub fn read(self) -> Result<usize, WaylandError>

Attempt to read events from the Wayland socket

If multiple threads have a live reading guard, this method will block until all of them are either dropped or have their read() method invoked, at which point on of the threads will read events from the socket and invoke the callbacks for the received events. All threads will then resume their execution.

This returns the number of dispatched events, or 0 if an other thread handled the dispatching. If no events are available to read from the socket, this returns a WouldBlock IO error.

Trait Implementations

impl Debug for ReadEventsGuard

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.