refactor(audio): share AudioHandler between iOS and macOS, bump deps

crates/chanora_audio/src/engine.rs: drop the macOS-specific event-queue producer/consumer path; macOS now uses the iOS-style direct AudioHandler::fill_buffer in the VPIO render callback. The shared AudioHandler is an Arc<Mutex<...>>; the realtime callback uses try_lock so it never blocks on the tokio decode task (see ios_voice_unit.rs render callback).

crates/chanora_audio/src/mobile_voice_backend.rs: update VoiceAudioParams cfg gates — handler is now the iOS/macOS/desktop shape (Arc<Mutex<AudioHandler<SessionAudioId>>>), event_producer is Android-only.

crates/chanora_audio/src/lib.rs: widen the audio_event_queue module visibility to test so the macOS-specific path can be exercised by the unit test suite.

Cargo.toml: bump cpal 0.17.3 -> 0.18.0, jni 0.21 -> 0.22.4, windows 0.54 -> 0.62, criterion 0.5 -> 0.8. Cargo.lock follows.

crates/chanora_audio/Cargo.toml

@@ -35,7 +35,7 @@ crossbeam = { version = "0.8", default-features = false, features = ["alloc", "c
 # Desktop audio I/O for Windows capture/playback and Linux capture.
 # Linux playback uses SDL2; Apple platforms use direct VoiceProcessingIO
 # AudioUnits via `coreaudio-rs` for the voice path.
-cpal = "0.17.3"
+cpal = "0.18.0"
 
 [target.'cfg(any(target_os = "ios", target_os = "macos"))'.dependencies]
 # Direct CoreAudio AudioUnit access on Apple platforms (DEC-011 follow-up).
@@ -58,7 +58,7 @@ ort = { version = "2.0.0-rc.12", default-features = false, features = ["load-dyn
 # JNI bindings to flip Android's AudioManager into MODE_IN_COMMUNICATION
 # when the voice-comm preset is requested. ndk_context is initialised
 # by the bridge crate's android_init shim.
-jni = { version = "0.21", default-features = false }
+jni = { version = "0.22.4", default-features = false }
 ndk-context = "0.1"
 # Oboe-rs (Google Oboe wrapper) for low-latency voice capture + playback.
 # Primary backend for SDD-111..SDD-115. The pre-compiled static library
@@ -84,7 +84,7 @@ bytemuck = { version = "1", features = ["derive"] }
 # + WM_INPUT translation backed by a hidden message-only window, and
 # SetWindowsHookExW(WH_KEYBOARD_LL / WH_MOUSE_LL) fallback. Both
 # require a per-backend OS thread that owns a message pump.
-windows = { version = "0.54", features = [
+windows = { version = "0.62", features = [
     "Win32_Foundation",
     "Win32_Graphics_Gdi",
     "Win32_System_LibraryLoader",
@@ -104,7 +104,7 @@ tracing-subscriber = { version = "0.3", features = ["registry"] }
 # SDD-120 §3 — criterion bench harness (realtime_capture / opus_codec /
 # resampler). `harness = false` per bench entry below disables the
 # default libtest harness so criterion can install its own.
-criterion = "0.5"
+criterion = "0.8"
 # SDD-120 §3 item 1 — dhat is used as the global allocator inside
 # `benches/realtime_capture.rs` to count post-warmup heap allocations
 # on the realtime capture path. Dev-dep only — does NOT affect
@@ -147,7 +147,7 @@ futures-util = { version = "0.3", default-features = false, features = ["std"] }
 # Random token bytes for the portal handle_token / session_handle_token
 # options. The portal recommends fresh tokens to scope its own
 # object paths per call.
-rand = "0.8"
+rand = "0.10"
 # SDL2 audio for Linux. Replaces the cpal playback path on Linux only;
 # cpal stays in use for Linux capture and Windows capture/playback.
 # Apple platforms use direct VoiceProcessingIO AudioUnits. Rationale: the
@@ -166,4 +166,4 @@ rand = "0.8"
 # libSDL2.so. Arch ships `sdl2-compat`; Debian/Ubuntu ship
 # `libsdl2-2.0-0`; Fedora ships `SDL2`. The chanora-flutter Linux
 # build documentation lists this as a runtime dependency.
-sdl2 = { version = "0.37", default-features = false }
+sdl2 = { version = "0.38", default-features = false }