fix(ci): cap cargo to two parallel jobs so concurrent test-binary links fit runner memory

Runs 514/516/519 (incl. master itself) died on 'ld terminated with
signal 9' — the workspace now links enough large test binaries that
per-core parallel linking OOMs the runner even at line-tables-only
debuginfo. CARGO_BUILD_JOBS=2 keeps at most two links in flight.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

.gitea/workflows/test.yml

@@ -35,8 +35,15 @@ jobs:
     # runner's memory — ld was OOM-killed (signal 9) on runs 447 and 486.
     # line-tables-only keeps file:line in panic backtraces while cutting the
     # link's memory footprint enough to fit the runner.
+    #
+    # CARGO_BUILD_JOBS=2: with one job per core, cargo links several large
+    # test binaries concurrently; as the workspace grew (runs 514/516/519)
+    # two+ simultaneous ld processes OOM-killed the runner again even at
+    # line-tables-only. Two jobs keeps at most two links in flight — the
+    # compile-throughput cost is small next to the cache-warm build.
     env:
       CARGO_PROFILE_DEV_DEBUG: line-tables-only
+      CARGO_BUILD_JOBS: '2'
 
     steps:
       - name: Checkout

solitaire_engine/src/safe_area.rs

@@ -82,7 +82,8 @@ impl Plugin for SafeAreaInsetsPlugin {
             );
 
         #[cfg(target_os = "android")]
-        app.init_resource::<android::SafeAreaPollTries>()
+        app.add_message::<crate::events::StateChangedEvent>()
+            .init_resource::<android::SafeAreaPollTries>()
             .add_systems(Update, android::refresh_insets)
             .add_systems(Update, android::rearm_on_resumed)
             .add_systems(Update, android::refresh_surface_size);
@@ -330,11 +331,16 @@ mod android {
     ///    `on_window_resized` in `table_plugin` recomputes the board layout,
     /// 3. re-arms the inset poller, because a screen change almost always
     ///    moves the system bars too — covering the "re-poll never fires"
-    ///    hole left open in #116.
+    ///    hole left open in #116,
+    /// 4. emits `StateChangedEvent` so `card_plugin`'s sync pipeline
+    ///    re-renders every card sprite from scratch — belt-and-braces for
+    ///    the transient tableau clip of #130, where geometry converged but
+    ///    stale card visuals survived the relayout.
     pub(super) fn refresh_surface_size(
         mut frame: Local<u32>,
         mut windows: Query<(Entity, &mut Window)>,
         mut resize_events: MessageWriter<WindowResized>,
+        mut state_events: MessageWriter<crate::events::StateChangedEvent>,
         mut poll: ResMut<SafeAreaPollTries>,
     ) {
         const POLL_INTERVAL_FRAMES: u32 = 30; // ~0.5 s @ 60 fps
@@ -381,6 +387,7 @@ mod android {
             width: decor_w as f32 / scale,
             height: decor_h as f32 / scale,
         });
+        state_events.write(crate::events::StateChangedEvent);
         poll.0 = 0;
     }