Struct SurfaceState 

struct SurfaceState {
    gl: Option<GLContext>,
    render: Option<MpvRenderContext>,
    mpv: *mut mpv_handle,
    frame_ready: Arc<AtomicBool>,
    callback_ctx: *mut Arc<AtomicBool>,
    logged_first_draw: bool,
    logged_first_frame: bool,
    logged_size: (i32, i32),
    logged_no_gl: bool,
    logged_no_window: bool,
    logged_no_size: bool,
    logged_render_fail: bool,
}

Everything the draw handler needs, shared with the GTK callbacks.

Fields

gl: Option<GLContext>

render: Option<MpvRenderContext>

mpv: *mut mpv_handle

frame_ready: Arc<AtomicBool>

Set by mpv’s update callback (on an mpv thread), cleared by the frame clock (on the main thread). The whole cross-thread contract.

callback_ctx: *mut Arc<AtomicBool>

The boxed clone of frame_ready handed to mpv, reclaimed on teardown. Null when no callback is registered.

logged_first_draw: bool

logged_first_frame: bool

logged_size: (i32, i32)

Last size we logged, so a size change re-reports rather than staying silent.

logged_no_gl: bool

logged_no_window: bool

logged_no_size: bool

logged_render_fail: bool

Implementations

impl SurfaceState

fn teardown(&mut self)

Tear down in the order DR-232 requires, with the GL context current.

The update callback is unregistered before the context is freed (inside destroy), and the GL objects go while their context is still bound. Getting this wrong is DR-184 on Android restated — a surface outliving its player — and is the likeliest cause of the one unexplained SIGSEGV the spike recorded.