In case the client isn't responding, this will block the compositor.
Instead we flush with MSG_DONTWAIT, which lets us fill up the kernel buffer
as much as we can (after not returning EPOLLOUT anymore it still can take
80k more), and then disconnect the client if we get EAGAIN.
In order to separate pointer and keyboard grabs, we need to
introduce a keyboard grab interface but first we must rename
some generic types to denote which device is holding the grab.
Type renames:
wl_grab_interface -> wl_pointer_grab_interface
wl_grab -> wl_pointer_grab
wl_input_device_start_grab -> wl_input_device_start_pointer_grab
wl_input_device_end_grab -> wl_input_device_end_pointer_grab
Remove the absolute coordinate fields from the pointer motion and
pointer_focus events. Clients are not supposed to see any global
coordinates.
Fix wayland-server code accordingly. wayland-client code is unaffected.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
The default grab implementation in wayland-server was updating the
focus resource before sending the button event. This would cause the
button release to be dropped from the implicit grab if the pointer is
moved away from the focus surface. This patch just swaps the order
around.
This commit changes the way struct wl_grab works in a couple of ways:
- The grab itself now decides when it ends instead of hardcoding button
up as the terminating event. We remove the end vfunc since a grab now
always know when it ends and can just clean up at that point.
- We add a new focus vfunc that is invoked every time the pointer enters
a new surface, regardless of any grabs. The callback receives the
surface and the surface-relative pointer coordinates. The callback lets
a grab send enter/leave events and change the grab focus.
- The grab has a focus surface, wich determines the coordinate space
for the motion callback coordinates.
- The input device always tracks the current surface, ie the surface that
currently contains the pointer, and coordinates relative to that surface.
With these changes, we will be able to pull the core input event delivery
and the drag and drop grab into the core wayland-server library.
Add a clean-up function for destroying all objects created in
wl_input_device_init(). Can be used to fix memory leaks reported by
Valgrind in the demos.
The init function was also missing an explicit initialisation of the
'keys' array. Add the explicit array init, although it is redundant with
the zeroing of the whole struct.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
krh: Edited to rename function to *_release()
On wl_display_add_socket(), the listening socket fd is added to the
event loop. However, wl_event_source object is not stored and hence
cannot be freed, resulting in a minor leak.
Store wl_event_source pointer in struct wl_socket so we can track it,
and destroy it on wl_display_destroy(). The event loop itself must be
destroyed after destroying the event sources linked to it.
Fixes a Valgrind reported memory leak.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
During client tear-down, all objects are destroyed in id order.
Therefore the display object is destroyed first.
If the destroy listeners of any object destroy another object by calling
wl_resoruce_destroy(), we try to send a delete_id event to the client.
This leads to a segmentation fault without a display object.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
This commit brings a big change to the DND and copy/paste interfaces.
Most importantly the functionality is now independent of wl_shell.
The wl_shell interface is intended for desktop style UI interaction and
an optional and experimental interface.
The new interface also allows receiving the DND data multiple times or
multiple times during the drag, and the mechanism for offering and receiving
data is now shared between DND and selections.
We set aside a range of the object ID space for use by the server. This
allows the server to bind an object to an ID for a client and pass that
object to the client. The client can use the object immediately and the
server can emit events to the object immdiately.
Some events, such as the display.delete_id, aren't very urgent and we
would like to not always send them immdiately and cause an unnecessary
context switch. The wl_resource_queue_event() function will place the
event in the connection output buffer but not request the main loop to
poll for writable. The effect is that the event will just sit in the
output buffer until a more important event comes around and requires
flushing.
We need to make sure the client doesn't reuse an object ID until the
server has seen the destroy request. When a client destroys an ID
the server will now respond with the display.delete_id event, which lets
the client block reuse until it receives the event.
We no long track the focused surface, but expect the compositor to set
focus when the focus changes. We do track the resource for the input
device that corresponds to the current surface, in case that goes away.
The visual interface was meant to be a generic mechanism for
specifying the content of a buffer. It goes back to before we had the
buffer factory interfaces (like wl_drm and wl_shm) and we wanted to
keep it open-ended enough that yuv, png or even svg buffer or so would
be possible.
Now that we have the buffer abstraction, we can add different buffer
types by introducing new interfaces that create buffers. It only
makes sense to leave it to those interfaces to specify the contents of
the buffers.
For wl_shm, this means that we now just specify the pixel format using
an enum. For EGL buffers, the exact pixel formats are controlled by
the implementation (part of wl_drm and similar), and from the client
point of view, everything is controlled using EGLConfigs.
So obvious in retrospect. The object system can do all the work for us
and keep track of pending calls as regular objects and we don't need to
abuse the resource system to get them cleaned up on client exit. We
don't need the custom key management or (broken) lookup, we just sue
object IDs. And last but not least, anybody can receive the callback,
not just display listeners.
