~raof/mir/wayland-clients-close-their-sessions-on-quit : changes

~raof/mir/wayland-clients-close-their-sessions-on-quit » Changes from revision 4278

From Revision 4278 to 4259

Rev	Summary	Authors	Date
4278	Wayland: Close client mf::Session on wl_client destruction. Wayland: Close client mf::Session on wl_client destruction.	Christopher James Ha...	6 years ago
4277	WlSurface: Don't try to send frame notifications after the W... WlSurface: Don't try to send frame notifications after the WlSurface has been destroyed. Like everything else we punt onto the wl_event_loop we need to ensure that the object is still live before executing the callback.	Christopher James Ha...	6 years ago
4276	Wayland: Handle WlPointer and WlTouch lifetimes in the same ... Wayland: Handle WlPointer and WlTouch lifetimes in the same way as WlKeyboard	Christopher James Ha...	6 years ago
4275	Wayland: Collect WlKeyboard lifetime handling into a run_unl... Wayland: Collect WlKeyboard lifetime handling into a run_unless wrapper	Christopher James Ha...	6 years ago
4274	Merge mir::client::Event branch Merge mir::client::Event branch	Christopher James Ha...	6 years ago
4273	Wayland: Add missed locking from new InputCtx::handle_event ... Wayland: Add missed locking from new InputCtx::handle_event override	Christopher James Ha...	6 years ago
4272	Wayland: Actually shrink the inputCtx::listeners vector when... Wayland: Actually shrink the inputCtx::listeners vector when removing listeners	Christopher James Ha...	6 years ago
4271	Wayland: Don't double-free WlSurfaces. Wayland: Don't double-free WlSurfaces. The lifetime handling for wayland::* classes requires that the destructors only be called via the destruction of the associated wl_resource.	Christopher James Ha...	6 years ago
4270	Wayland: Handle keyboard lifetime, take 2. Wayland: Handle keyboard lifetime, take 2. Unfortunately, keyboard.release() is not the only way that keyboards can get destroyed. All client resources will be destroyed on client disconnect, so ~WlKeyboard can't rely on WlKeyboard::release() having ensured that no events are queued to the keyboard on the Wayland event loop. Instead maintain a “destroyed” flag to short-circuit the event handler callbacks once the WlKeyboard has been destroyed.	Christopher James Ha...	6 years ago
4269	Wayland: Lifetimes should be tied to client resources only f... Wayland: Lifetimes should be tied to client resources only for non-global objects. Globals are not tied to client lifetime - the globals themselves persist as long as the shell wants them to. There are client resources associated on bind, but these are separate from the global resource. Instead, properly set up non-global destructors from their constructors.	Christopher James Ha...	6 years ago
4268	Wayland: Use evdev codes for pointer buttons Wayland: Use evdev codes for pointer buttons	Christopher James Ha...	6 years ago
4267	Merge trunk Merge trunk	Christopher James Ha...	6 years ago
4266	Wayland: Handle wl_{pointer,keyboard,touch} lifetime properl... Wayland: Handle wl_{pointer,keyboard,touch} lifetime properly.	Christopher James Ha...	6 years ago
4265	Wayland: Tie the lifetime of wrapper classes to the lifetime... Wayland: Tie the lifetime of wrapper classes to the lifetime of the wl_resource. This essentially turns the underlying wl_resource into a unique_ptr<> for our wrapper classes. Thus, when the resource goes away our destructor is called, and we must ensure that the destructor is called only when the resource goes away.	Christopher James Ha...	6 years ago
4264	Wayland: Protect the set of Wl* listeners behind a lock. Wayland: Protect the set of Wl* listeners behind a lock. register_listener/unregister_listener will be called on the Wayland event loop, but handle_event will be called from whatever thread the server (or shell!) chooses.	Christopher James Ha...	6 years ago
4263	Wayland: the protocol now says that wl_pointer buttons are d... Wayland: the protocol now says that wl_pointer buttons are defined by evdev	Christopher James Ha...	6 years ago
4262	Wayland: Don't crash when we get an unknown pointer button. Wayland: Don't crash when we get an unknown pointer button.	Christopher James Ha...	6 years ago
4261	mgm::BufferAllocator: Handle a wl_buffer being attached mult... mgm::BufferAllocator: Handle a wl_buffer being attached multiple times. As with the WlShmBuffer, it's valid for a wl_buffer to be wl_surface.attach()ed to multiple wl_surface-s at the same time. If a buffer is attached twice then the buffer.release() event needs to be sent exactly once, when the compositor is finished with all submissions. So, do what we do with WlShmBuffer - associate a shared_ptr<WaylandBuffer> with each wl_buffer, and hand out references to it rather than creating a new WaylandBuffer on each (committed) attach(). . Approved by mir-ci-bot, Gerry Boland, Alan Griffiths.	Christopher James Ha...	6 years ago
4260	Migrate ClientSurfaces.* tests to miral APIs. Migrate ClientSurfaces.* tests to miral APIs. Approved by Gerry Boland, mir-ci-bot.	Alan Griffiths	6 years ago
4259	[miral-kiosk] Parentless freestyle windows should also be ma... [miral-kiosk] Parentless freestyle windows should also be maximized. Approved by mir-ci-bot, Gerry Boland.	Alan Griffiths	6 years ago

Older »