punktfunk/clients/apple

punktfunk Apple client (SwiftUI)

The native macOS/iOS client for punktfunk/1 (the post-GameStream protocol). All networking/protocol work — QUIC control plane, UDP data plane, GF(2¹⁶) FEC, AES-GCM, input datagrams, Opus audio, cert pinning — lives in the shared Rust core (statically linked as PunktfunkCore.xcframework); this package is the Swift shell: decode (VideoToolbox), present (SwiftUI), input capture.

Status — first light achieved (2026-06-10)

Validated live, Mac ↔ Linux box over the LAN: gamescope virtual output → NVENC HEVC → punktfunk/1 (GF(2¹⁶) FEC + AES-GCM over UDP, QUIC control) → VideoToolbox → AVSampleBufferDisplayLayer on glass at 1280×720@60, with mouse/keyboard flowing back as QUIC datagrams into the host's gamescope EIS injector (thousands of events injected during the session). Headless variant of the same proof: RemoteFirstLightTests decoded 60/60 received AUs spanning 983 ms of host capture clock.

The connector underneath (punktfunk_core::client::NativeClient over the C ABI) carries the full session: video AUs, Opus audio (nextAudio()), rumble (nextRumble()), input incl. gamepads, and cert pinning + TOFU (pinSHA256:/hostFingerprint) — see m3.rs::tests::c_abi_connection_roundtrip (three sequential sessions: TOFU, pinned reconnect, wrong-pin rejection). The host (punktfunk-host m3-host) is a persistent listener: reconnect at will during development.

What's here, all compiled and tested on macOS (Xcode 26.5 / Swift 6.3):

• PunktfunkKit (library)
  • PunktfunkConnection.swift — wrapper over the C ABI. AUs/audio are copied into Data (the C pointer is only valid until the next call of the same kind). close() is safe from any thread: per-plane locks enforce the C contract ("never close with a next_au/next_audio in flight") instead of leaving it to callers. Pinning + TOFU via pinSHA256:/hostFingerprint.
  • AnnexB.swift — in-band VPS/SPS/PPS → CMVideoFormatDescription; Annex-B → AVCC CMSampleBuffer with DisplayImmediately set.
  • StreamView.swift — SwiftUI NSViewRepresentable over AVSampleBufferDisplayLayer (stage-1 presenter: the layer hardware-decodes compressed HEVC itself). One pump thread per view, token-cancelled so reconnects can't double-pump.
  • InputCapture.swift — GCMouse raw deltas + GCKeyboard HID→VK mapping (the host's vk_to_evdev consumes Windows VKs), with fractional-delta accumulation so sub-pixel motion isn't truncated away. Buttons use GameStream ids (1=left … 5=X2); scroll is WHEEL_DELTA(120)-scaled.
• PunktfunkClient (the app): hosts grid (saved in UserDefaults), "+" toolbar sheet to add hosts, stream mode in Settings (⌘,), two trust flows — the trust-on-first-use fingerprint prompt over the live-but-blurred stream, and SPAKE2 PIN pairing (PairSheet, from a host card's context menu or the trust prompt; ClientIdentityStore keeps the client identity in the Keychain and presents it on every connect) — then pinned reconnects, fps/Mb-s HUD. (Audio playback and gamepad capture are not wired into the app yet — the connector surface is there; see notes 5–6.)
• Tests (swift test): byte-level Annex-B units; a real-codec round trip (VTCompressionSession-encoded HEVC rebuilt as the host's wire shape → AnnexB → VTDecompressionSession → pixels); loopback integration against real local hosts (test-loopback.sh — stream round trip, plus the PIN pairing ceremony and the --require-pairing gate against a second, armed host); the remote first-light test above.

Build / run / test (on a Mac)

rustup target add aarch64-apple-darwin x86_64-apple-darwin
bash scripts/build-xcframework.sh        # → clients/apple/PunktfunkCore.xcframework
cd clients/apple
swift build && swift test                # loopback/remote tests self-skip without a host
swift run PunktfunkClient                # the unbundled dev shell (CLI)
open Punktfunk.xcodeproj                 # the real app: ⌘R builds + runs Punktfunk.app

bash test-loopback.sh                    # full loopback proof: builds punktfunk-host
                                         # (synthetic source — runs on macOS), streams
                                         # byte-verified frames into the Swift client

# against the real host (Linux box, see CLAUDE.md "Running on this box") — m3-host is a
# persistent listener, reconnect at will:
#   PUNKTFUNK_COMPOSITOR=gamescope PUNKTFUNK_GAMESCOPE_APP=vkcube PUNKTFUNK_ZEROCOPY=1 \
#   cargo run -rp punktfunk-host -- m3-host --source virtual --seconds 60
PUNKTFUNK_REMOTE_HOST=<box-ip> swift test --filter RemoteFirstLightTests   # headless
PUNKTFUNK_AUTOCONNECT=<box-ip> PUNKTFUNK_MODE=1280x720x60 swift run PunktfunkClient # on glass

Xcode project (Punktfunk.xcodeproj)

The app target Punktfunk wraps the same sources as the swift run shell (Sources/PunktfunkClient, a synchronized folder — no duplication) plus App/ (asset catalog) and links PunktfunkKit from the local package. Generated Info.plist, ad-hoc signing, bundle id io.unom.punktfunk. Notes:

• App icon: App/Assets.xcassets ships an empty AppIcon slot. For an Icon Composer .icon: add the file to the project (target Punktfunk), set it as the App Icon in the target's General tab, and delete the placeholder AppIcon.appiconset. Heads-up: CLI actool (Xcode 26.5) crashed compiling punktfunk_Logo.icon — if Xcode does the same, suspect the icon bundle (it has a duplicate-named layer, "…Layer-3 2.svg"), not the project.
• Tests from Xcode: the package tests run with swift test; to get them on ⌘U, add PunktfunkKitTests once via Edit Scheme → Test → + (Xcode persists it into the shared scheme — a hand-written package-test reference doesn't resolve headlessly).
• xcodebuild -project Punktfunk.xcodeproj -scheme Punktfunk build works headlessly.

Notes for whoever picks this up next

1. cbindgen import quirk (the predicted "small compile fixes", now fixed): the C17-compatible header spells PunktfunkStatus/PunktfunkInputKind as integer typedefs while the enum constants import into Swift as a distinct same-named type — bridge with .rawValue (see the top of PunktfunkConnection.swift). Don't fight the generated header.
2. ABI contract: one video pump thread per connection, plus optionally one separate audio drain thread for nextAudio()/nextRumble() (the core keeps per-plane borrow slots, so the planes never alias); send() is enqueue-only and safe alongside all of them. The wrapper's per-plane locks make close() safe from anywhere (it waits out in-flight polls, ≤ their timeouts).
3. Decode flow: the host opens every stream with an IDR carrying VPS/SPS/PPS in-band and recovery keyframes re-send them — "refresh the format description on every IDR" (what StreamView does) is sufficient; there is no out-of-band extradata, ever.
4. Stage 2 (next): explicit VTDecompressionSession + CAMetalLayer for frame-pacing control (ProMotion/120 Hz), glass-to-glass measurement via tools/latency-probe (the host stamps pts_ns with its capture wall clock; across machines you need a clock offset estimate from the QUIC RTT).
5. Audio: nextAudio() yields raw Opus packets (48 kHz stereo, one 5 ms frame each, sequence-numbered). Decode with libopus or AVAudioConverter/kAudioFormatOpus into an AVAudioEngine source node; conceal gaps (drop/dup) rather than blocking — the Rust side buffers 320 ms and drops the newest packet when the puller lags. Wall-clock ptsNs shares the host clock with video AUs for A/V sync. Wiring this into PunktfunkClient is the next app-side task.
6. Gamepads: GCController → .gamepadButton(...)/.gamepadAxis(...) events (wire contract documented on the constructors; the host accumulates them into a virtual Xbox 360 pad). Poll nextRumble() and feed GCDeviceHaptics for force feedback. Client-side capture isn't in InputCapture yet.
7. Trust — the full ceremony exists now (SPAKE2). generateIdentity() once (persist both PEMs in the Keychain), then pair(host:identity:pin:name:) with the 4-digit PIN the host prints when it ARMS pairing (--allow-pairing/--require-pairing; one PIN per arming window, shown at startup — the user reads it before pairing). Returns the host's VERIFIED fingerprint; persist it and pass pinSHA256: + identity: to every connect. Pairing is a real PAKE: a wrong PIN gets ONE online guess (no offline dictionary attack), throwing .wrongPIN; a wrong-size pin throws .invalidPin. PunktfunkClient implements both flows: the TOFU fingerprint sheet keeps working against hosts not running --require-pairing, and the PIN ceremony is wired in — ClientIdentityStore (Keychain) on every connect, PairSheet from a host card's context menu or the trust prompt's "Pair with PIN instead…" (the host's accept loop is sequential, so that path drops the live session before pairing). With --require-pairing the host now authorizes clients too (the "other direction" is no longer open, opt-in per host); the whole gate is regression-tested in testPairingCeremonyAndRequirePairingGate. 7b. Resize without reconnect: requestMode(width:height:refreshHz:) mid-stream — the host rebuilds at the new mode in ~90 ms; the first new-mode AU is an IDR with fresh parameter sets (the refresh-on-IDR decode flow handles it untouched) and currentMode() reflects the switch. Wire it to window-resize events.
8. Input capture caveats (stage 1): GC handlers only fire while the app has focus — on focus loss InputCapture auto-releases everything still held (keys + buttons) so nothing sticks down host-side. While the stream has focus the LOCAL cursor is hidden and frozen mid-view (CursorCapture in StreamView.swift — the host renders its own cursor; the local one diverges from it and a stray click would focus another app); Cmd+Tab frees it, ⌘D disconnects. While captured, key NSEvents are swallowed by a local event monitor (GC reads HID directly; without it every keystroke bubbles up the responder chain unhandled and NSWindow beeps) — except ⌘-combos, which still reach the local app (⌘D/⌘Q) in addition to the host; a capture toggle is a small follow-up. One live capture per process (the GC mouse/keyboard singletons have a single handler slot — ownership is tracked so a stale capture's stop() can't clobber a newer one).
9. iOS: same package (BUILD_IOS=1 for the xcframework slice); StreamView needs the UIViewRepresentable twin and touch→input mapping.

Known limitations of the current host (relevant to client UX)

• One session at a time (the listener is persistent, but a second concurrent client waits in the accept queue until the current session ends — the virtual output and encoder are single-tenant).
• Mid-stream renegotiation (resolution change without reconnect) is designed-for but not implemented (the Welcome is one-shot today).
• Host-side gamepad injection needs /dev/uinput access on the box (udev rule from docs/linux-setup.md).