Address review feedback: fix comment formatting and consolidate duplicate Logger.Log calls

Agent-Logs-Url: https://github.com/winnerspiros/osu/sessions/3c1fa36e-ab75-466a-9437-b218eeb9fa8e

Co-authored-by: winnerspiros <1675249+winnerspiros@users.noreply.github.com>

Author: Copilot

osu.Android/OsuGameActivity.cs

@@ -683,20 +683,16 @@ public void SurfaceChanged(ISurfaceHolder holder, global::Android.Graphics.Forma
             // picks up the new ANativeWindow and negotiates a proper BGRA/RGBA 8-bit swapchain.
             if (format == global::Android.Graphics.Format.Rgb565 && LogManagement.IsVulkanConfigured())
             {
-                Logger.Log(
-                    "[osu!] Android surface pixel format RGB565 detected mid-session (Vulkan path). " +
-                    "Requesting RGBA8888 and triggering a surface recreate. " +
-                    "If this fires after startup it indicates an OEM display-mode change " +
-                    "(e.g. SetSustainedPerformanceMode) reset the surface format.",
-                    LoggingTarget.Runtime,
-                    LogLevel.Important);
-                Logger.Log(
-                    "[osu!] Android surface pixel format RGB565 is incompatible with the Vulkan rendering pipeline " +
-                    "— requesting RGBA8888 and triggering a surface recreate. " +
-                    "This is the root cause of the Vulkan black-screen crash on Adreno (SDL_PIXELFORMAT_RGB565 in runtime log). " +
-                    "The next SurfaceChanged will carry the corrected format.",
-                    LoggingTarget.Performance,
-                    LogLevel.Important);
+                // Log to Runtime so the mid-session RGB565 reset is visible in the main log
+                // (and therefore in the notification overlay). Performance log gets the same
+                // entry for correlation with display-mode and frame-timing data.
+                string rgb565Message =
+                    "[osu!] Android surface pixel format RGB565 detected (Vulkan path) — " +
+                    "requesting RGBA8888 and triggering a surface recreate. " +
+                    "If this fires after startup an OEM display-mode change reset the surface format, " +
+                    "which would cause a mid-session swapchain rebuild at wrong dimensions.";
+                Logger.Log(rgb565Message, LoggingTarget.Runtime, LogLevel.Important);
+                Logger.Log(rgb565Message, LoggingTarget.Performance, LogLevel.Important);
 
                 try
                 {

osu.Android/OsuGameAndroid.cs

@@ -379,8 +379,8 @@ protected override void LoadComplete()
             //
             // Pinning Update + Draw + Input to a 5-core subset (mask 0xF8 on SD8G2) is the
             // ONLY unconditional Android-specific synchronous mutation we still perform
-            // during the cold-start window — every other customisation (
-            // RequestUnbufferedDispatch, refresh-rate selection, Oboe / Vulkan-probe init,
+            // during the cold-start window — every other customisation (RequestUnbufferedDispatch,
+            // refresh-rate selection, Oboe / Vulkan-probe init,
             // performance-mode GC-latency flip) is already deferred behind the
             // refreshRateDelayMs scheduler below. Field logs.zip on v2026.423.176 show both
             // a normal launch and a safe-mode launch dying silently mid-Toolbar load
@@ -1019,8 +1019,7 @@ protected override void LoadComplete()
             // the BDL load thread, in the silent cold-start window — exactly
             // when we are debugging a startup hang. Deferring the initial
             // fire via Scheduler.AddDelayed onto the same refreshRateDelayMs
-            // timer that gates the initial refresh-
-            // rate apply / the initial performance-mode apply keeps the cold-
+            // timer that gates the initial refresh-rate apply / performance-mode apply keeps the cold-
             // start path free of synchronous native init even when a saved-
             // true setting would otherwise force it, AND ensures the native
             // init actually lands AFTER the cold-start Toolbar texture-upload