Polling the event queue before NSApp had been allowed to finish
launching, in our case by starting our self-terminating run loop,
triggered an assertion inside NSApplication.
This fix, which makes all event processing functions capable of starting
it, makes that assertion less likely.
A more Cocoa-friendly fix would be to finish launching NSApp during
glfwInit and let people annoyed by the menu bar disabled it with
GLFW_COCOA_MENUBAR. That may not be suitable for 3.3-stable, though.
Fixes#1543.
This lets compositors avoid re-rendering the entire buffer when only the
outside of the squares changed.
If glfwGetBufferAge() returns 0 for any reason (the buffer was just
created, there was an error, or the underlying API doesn’t track buffer
age), we swap the entire buffer again.
This new API lets the application know when the current buffer was last
written to. If it returns 0, it means the user shouldn’t rely on what
was last present in this buffer and should draw everything again, as
usual.
This provides a significant boost in efficiency, especially on tiling
GPUs, by letting the application avoid to clear and redraw everything at
every frame.
wl_surface supports a damage_buffer request since its version 4, which
requires wl_compositor to have been bound at that version too.
damage_buffer can then be used by the EGL implementation to optimise
eglSwapBuffersWithDamageKHR().
This new API helps the user’s compositor reduce memory bandwidth and
thus power usage by only re-rendering a bunch of dirty rectangles
instead of the entire application buffer.
There is no guarantee that it will effectively get used, it is perfectly
valid to continue damaging the entire buffer instead like with
glfwSwapBuffers() in some cases.
That way the application only sees the cursor moving when it is inside
of its area, it won’t go back to the top or left side when trying to
resize the window or just hovering the fallback decorations.
Previously, any pointer motion in the window decorations when using the
fallback implementation would obtain the wl_cursor again, and do the
attach danse for no benefit.
This will ultimately allow animated cursors to not reset to the first
frame on motion, once these will be implemented.
This falls back to calculating the monitor physical size from the
current resolution and the default X11 DPI when the physical size
returned by RandR is zero.
