Phase 1 of design/pyrowave-444-hdr.md. No behavior change yet: the handshake's
4:4:4 gate now admits PyroWave (probe = can_encode_444(codec), capture gate
inherently satisfied — the wavelet path always ingests an RGB source and does
its own CSC), but can_encode_444 stays false for PyroWave until the per-OS
full-res-chroma CSC variants land (Phase 2 Linux, Phase 3 Windows), so every
session still resolves 4:2:0/8-bit.
- Both host encoders take the negotiated ChromaFormat (bail on 444 for now);
the PUNKTFUNK_ENCODER=pyrowave lab override pins 4:2:0.
- Bitrate: the automatic ~1.6 bpp pin resolves AFTER depth+chroma and scales
x1.625 for 4:4:4 / x1.15 for 10-bit (factors from the Phase-0 fixture
matrix); the mid-stream mode-switch re-resolve threads the session's values.
- Client: PyroWaveDecoder builds its plane ring (full-res chroma when 444) and
creates the upstream decoder from the negotiated chroma, keeps chroma fixed
across mid-stream resizes, drops the even-dims requirement for 444, and
returns the negotiated Welcome ColorInfo as the frame colour contract
instead of hardcoded BT.709 (the wavelet bitstream has no VUI).
Verified on .21 (RTX 5070 Ti): clippy -D warnings (host+client+encode), host
186 tests, client + pf-encode tests, fmt, and the pyrowave_smoke GPU
round-trip through the patched vendored lib (97cf15e3).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Upstream sizes the raw quantized-payload worst-case buffer at
aligned_width*aligned_height*2 bytes: real headroom for 4:2:0's 1.5
samples/px, but half of what 4:4:4's 3 samples/px can produce. Busy 4:4:4
content overruns the buffer on the GPU and corrupts the adjacent meta/bucket
suballocations — nondeterministic corrupt bitstreams and encoder crashes at
ANY target bitrate (smooth content never trips it, which is why 4:2:0 and
simple 4:4:4 both look fine). Found by the Phase-0 measurement matrix for
design/pyrowave-444-hdr.md; fix validated alone via upstream's own devel
tools on the RTX 5070 Ti: deterministic byte-identical outputs across runs,
1080p + 4K, 8- and 16-bit, PSNR at the expected operating points.
Patch lives in crates/pyrowave-sys/patches/ and vendor-pyrowave.sh now
re-applies patches on re-vendor; PUNKTFUNK-VENDOR.txt records it. Upstream
report to follow. No wire/ABI change; 4:2:0 sizing unchanged.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The runner only found unscoped `punktfunk-plugin-*` packages. An unscoped
package can't be split across registries, which forced plugins to bundle their
own copy of @punktfunk/host + effect (the plugin's Gitea registry has no
`effect`; `--registry <gitea>` 404s it). Discovering the scoped
`@punktfunk/plugin-*` convention lets a plugin resolve cleanly from one scope
map and depend on @punktfunk/host + effect as SHARED (hoisted) deps instead of
bundling — no per-plugin duplication. Additive: unscoped names still work.
Bumps @punktfunk/host to 0.1.1.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The DCP "mismatched swapID's" kernel panic reproduced on a 240 Hz Mac
Studio with stage-3 glass pacing active: a fully serialized,
one-in-flight present stream still races WindowServer's own swap
submissions. So the mitigation has to change the MECHANISM, not the
rate — the CAMetalLayer image queue itself is the racing path in a
composited (windowed) session.
Windowed PyroWave now presents the way video players do: the planar
CSC renders into a pooled IOSurface (4 × BGRA8, in-use-aware LRU
reuse) and the render thread hands it to a plain CALayer's `contents`
on main inside an ordinary CATransaction. WindowServer treats that as
normal layer damage on its own composite cadence — no out-of-band
image-queue swaps to race. Fullscreen keeps the CAMetalLayer path
(direct scanout, no compositing, no panic reports); the hosting view
pushes the window's composited state on every layout, and flipping
modes just covers/uncovers the metal layer (no black flash).
VT codecs keep the metal path everywhere: no panic reports there, and
their HDR/EDR presentation has no surface-contents equivalent wired.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
pyrowave's encoder fills the BitstreamSequenceHeader with `= {}` and its C API
offers no way to set colour/range, so it signals ycbcr_range=0=FULL — but both
host CSCs (rgb2yuv.comp on Linux, BgraToYuvPlanes on Windows) always emit BT.709
LIMITED Y'CbCr (black = Y'16). A client that honours the VUI (the Apple wavelet
decoder reads bit 30 of word1) then skips the limited→full expansion and shows
washed-out, raised blacks — reported on both Linux and Windows hosts.
Patch the range bit HONEST (mark_limited_range in the shared pyrowave_wire, called
by both encoders after packetize). Clients that hardcode limited (the Vulkan
video_pyrowave path) are unaffected, and pyrowave's own decode ignores the flag
(raw Y'CbCr reconstruction). No client rebuild needed. Unit-tested + asserted in
pyrowave_win_smoke; the smoke decode still round-trips 100/180/60.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Wire PyroWave into the Windows host (design/pyrowave-windows-host-zerocopy.md).
Before this a macOS client + Windows host that both selected PyroWave silently ran
HEVC: the host never advertised CODEC_PYROWAVE and open_video_backend bailed.
Approach (zero-copy, no GPU→CPU→GPU): pyrowave owns its own Vulkan device
(create_device_by_compat, by render-GPU vendor/device-id — NOT LUID, invalid in
Session 0). The capturer runs a BGRA→YUV BT.709-limited CSC (matching rgb2yuv.comp)
into TWO SEPARATE shareable plane textures — full-res R8 Y + half-res R8G8 CbCr —
which the encoder imports into pyrowave's device. Separate single/two-component
textures import reliably on NVIDIA at any size; a single planar NV12 import does NOT
(the vendored interop test: "only very specific resource sizes" — confirmed on-glass:
1024² fine, 720p/1080p/1440p garbage). A shared D3D11 fence, signalled after the CSC,
is imported as a Vulkan timeline semaphore so the wavelet read is ordered after it.
- pf-encode: enc/windows/pyrowave.rs (Encoder impl, two-plane import + Linux-style
plane views); host_wire_caps advertises CODEC_PYROWAVE on Windows when the backend
isn't Software; open_video_backend routes a negotiated PyroWave session first;
pyrowave-sys on the Windows target; interop confirmed at open → clean HEVC fallback.
- pf-encode: shared, unit-tested enc/pyrowave_wire.rs (single source of truth for the
client-facing AU framing); Linux encoder uses it too.
- pf-capture: dxgi.rs BgraToYuvPlanes CSC; idd_push.rs pyrowave mode — forces the
virtual display SDR (the VideoProcessor can't ingest the FP16 HDR ring), a
two-plane shareable out-ring, a shared fence passed every frame (so a rebuilt
encoder re-imports it). Threaded via OutputFormat::pyrowave.
- pf-frame: D3d11Frame::pyro carries the CbCr plane + fence; OutputFormat::pyrowave.
Verified on .173 (RTX 4090): full-host build + clippy -D warnings (nvenc,amf-qsv) +
fmt --all --check; pyrowave_wire unit tests; pyrowave_win_smoke GPU test round-trips
distinct Y/Cb/Cr (100/180/60) exactly at 1024²/720p/1080p/1440p; Stage-0 interop
validated in the real Session-0 service context on-glass. Deployed to the box.
Owed: final on-glass picture/latency confirmation.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The host .deb was built on the Ubuntu 26.04 rust-ci image, so dpkg-shlibdeps
baked in `Depends: libavcodec62` (FFmpeg 8) and a glibc-2.41 floor — making it
uninstallable on Ubuntu 24.04 LTS (FFmpeg 6.1 / libavcodec60, glibc 2.39; apt
reports the deps as "too recent"). The source floor (ffmpeg-next 8, libavcodec
>=61 APIs) means a straight 24.04 rebuild would fail too.
Build the host on Ubuntu 24.04 instead — lowering the glibc floor to 2.39 so one
binary runs on 24.04 -> 26.04 — and bundle a from-source LGPL FFmpeg 8 into the
package so it no longer depends on the distro libav*. Everything else the host
links is soname-compatible on 24.04 (opus is vendored via cmake; NVENC/libcuda
are dlopen-only, never link-time), and the only FFmpeg encoders used are
*_nvenc / *_vaapi (software H.264 fallback is the BSD-2 openh264 crate, not
FFmpeg libx264), so an LGPL build keeps the bundle license-clean.
- ci/rust-ci-noble.Dockerfile (new): ubuntu:24.04 builder; nv-codec-headers +
FFmpeg 8 (--enable-nvenc --enable-vaapi, shared) -> /opt/ffmpeg; PKG_CONFIG_PATH.
- packaging/debian/build-deb.sh: BUNDLE_FFMPEG=1 copies libav*/libsw*/libpostproc
into /usr/lib/punktfunk-host, patchelf-sets the rpath ($ORIGIN per-lib + binary
--force-rpath), feeds them to dpkg-shlibdeps (captures libva2/libdrm2), and drops
the libav* sonames from Depends. Normal (non-bundle) path unchanged.
- .gitea/workflows/deb.yml: split into build-publish (client/web/scripting on the
26.04 image) and build-publish-host (noble image, BUNDLE_FFMPEG=1); parallel,
identical version step, same apt distribution -> one universal host .deb.
- .gitea/workflows/docker.yml: build+push punktfunk-rust-ci-noble.
- packaging/debian/README.md: document the 24.04 LTS path + bundled local build.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Windowed PyroWave sessions were kernel-panicking Macs ("mismatched
swapID's" @UnifiedPipeline.cpp, WindowServer → AppleMobileDisp*-DCP).
The panic is an Apple DCP bug (SketchUp/Unity/240Hz-monitor reports hit
the same signature), but our stage-2 arrival pacing feeds the trigger
pattern: displaySyncEnabled=false out-of-band presents arriving faster
than the compositor latches them in a composited (windowed) session.
PyroWave makes that pattern routine — near-instant Metal wavelet decode
turns network clumps into same-millisecond present bursts, and it is
the codec that sustains stream rates above panel refresh.
Mitigation: PyroWave sessions on macOS now default to the stage-3
PresentGate (one presented-but-undisplayed swap in flight, serialized
on the on-glass callback, 100 ms stale backstop) — the racing pattern
cannot occur. Explicit stage2/stage3 picks (setting or
PUNKTFUNK_PRESENTER) still win, so the arrival-pacing A/B stays honest.
VideoToolbox codecs keep arrival pacing (decode latency spaces their
presents; no panic reports there).
PresenterChoice gains an `explicit` resolver (nil = no user selection)
so the codec-conditional default only applies when the user hasn't
picked a stage; pacing selection is a testable helper carrying the
rationale.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The "Which should I use?" table on the clients page listed most client
names as bold text, not links — only the Decky plugin and pf-webos were
clickable — so the client references appeared broken. Link each client to
its section (or dedicated page), and fix a stale Windows headless command.
Repo-wide docs correctness/staleness pass against the code:
- steam-deck: client-not-found -> flatpak-not-found (the real backend code)
- install: host cert is punktfunk-host-windows_<ver>.cer, not ..._setup.cer
- configuration: GPU_PRIORITY_CLASS default is auto; 10BIT/444 are default-on
- how-it-works/index: GameStream/Moonlight is opt-in (--gamestream)
- roadmap: clipboard sync is shipped, not planned
- install-client: MSIX/cert artifacts are arch-suffixed (_x64/_arm64)
- requirements: fix garbled 22H2/IddCx sentence
- status: Linux encode also covers AMD/Intel (VAAPI/Vulkan Video)
- automation: add the plugins.changed event
- windows-host: note the optional bundled VB-CABLE virtual mic
- sway: PUNKTFUNK_COMPOSITOR=hyprland is a wlroots-family alias
- running-as-a-service: punktfunk-probe is a source-build-only dev tool
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Regenerated api/openapi.json (host built on Linux) with the plugins tag —
registerPlugin/listPlugins/deregisterPlugin/getPluginUiCredential + the
PluginRegistration/PluginSummary/PluginUi/PluginUiPublic/UiCredential schemas —
and the SDK's generated client. Drift test + 44 mgmt/events host tests pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The clippy step ran in the default (debug) profile, compiling the whole
dep tree into a second target dir (C:\t\debug) — a second full build of
openh264-sys2's vendored C++ (pf-encode's software-H.264 fallback) on top
of the release copy the Build steps already produced. That second cc-rs
cl.exe fan-out tips the self-hosted runner into
`cabac_decoder.cpp: fatal error C1069 (cannot read compiler command line)`
— environmental disk/temp exhaustion, not a source error: the identical
file compiles fine in the release build minutes earlier, and openh264 is
untouched here. Linting in --release reuses the release build-script
artifacts (no openh264 rebuild) and keeps everything in one C:\t\release
tree. Matches pf-vkhdr-layer's clippy, which already runs --release.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The workspace went ring-only in 0.13.0 (aws-lc-sys breaks Windows CI), and
`ring` can sign with an existing RSA key but cannot *generate* one —
rcgen's ring backend returns `KeyGenerationUnavailable` for
`generate_for(&PKCS_RSA_SHA256)`. So `ServerIdentity::{load_or_create,
ephemeral}` panics with "rcgen RSA keygen / There is no support for
generating keys for the given algorithm" on any host without an existing
`cert.pem`. This is the shared trust root for both the GameStream TLS cert
and the QUIC identity clients pin, so a *fresh* 0.13.0 install can't start
the host at all (existing dev boxes survive only because they already have
a cert on disk); in CI every test that builds a `ServerIdentity` (all of
`mgmt::tests`, `gamestream::nvhttp`, and the `native::tests` synthetic host
via `serve()`) failed — 30 failures from this one cause.
Moonlight requires an RSA-2048 identity, so generate the key with the
pure-Rust `rsa` crate (already a dependency for the pairing signer) and
hand its PKCS#8 PEM to rcgen, whose ring backend *can* load and self-sign
with an externally supplied RSA key (`from_pkcs8_pem_and_sign_algo` →
`RsaKeyPair::from_pkcs8`). Return that same PEM so it stays byte-identical
to what `from_pems` re-parses.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`is_transient_io` only recognized `ENOBUFS` on unix; on Windows the
`#[cfg(not(unix))]` arm returned `false`, so `WSAENOBUFS` (10055) — which
Rust maps to `ErrorKind::Uncategorized`, missing the `WouldBlock` arm —
propagated out of the USO `send_gso` path through `Session::send_sealed`
and killed the session with `native::stream` "send failed — stopping
stream". A high-bitrate keyframe burst (one `WSASendMsg` USO super-buffer
is up to ~512 segments) momentarily exhausts the socket send buffer / AFD
non-paged pool; it's a lossy drop that FEC + the next frame recover,
exactly like the unix `ath11k` `ENOBUFS` case the classifier already
handles. Fires independently of client platform (Windows/macOS clients
both saw the crash) because it's the Windows *host's* send socket.
Add a `#[cfg(windows)]` arm matching `WSAENOBUFS` plus the
`WSAENET*`/`WSAEHOST*` network-path family (the Windows counterparts of
the droppable unix set), and extend the classifier test with per-platform
raw-errno coverage so the Windows CI runner exercises the 10055 path.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
docs/automation.md promises 'changes apply immediately, no restart' for BOTH
config paths, but the store loaded hooks.json once (OnceLock) — only PUT
/api/v1/hooks applied live, and a hand-edited file silently did nothing until
the next host start. Since the file is the accessible path most users will
take, make it real: get() re-stats the file per event (mtime + length) and
re-reads it on change, with the same lenient contract as startup (missing =
no hooks, invalid = disabled loudly). set() records the identity it wrote so
the API path doesn't trigger a spurious reload.
Validated live on the 0.13.0 host (.21): full connect/disconnect lifecycle
fires exec hooks (client.*/session.*/stream.*) — the reload gap was found
setting that test up.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Companion to the host ring fix: pf-client-core declared rustls with
features=["ring"] but WITHOUT default-features=false, so rustls's default
aws-lc-rs backend came along too (feature tree: aws-lc-rs <- rustls default <-
pf-client-core default) — pulling aws-lc-sys 0.41.0 into the Windows client +
session + presenter builds, the same C dep that fails on the CI runner. Pin
default-features=false + [ring, logging, std, tls12] (matching the tray/host/core
pattern). aws-lc-sys now gone from EVERY windows-built crate (host, client-windows,
client-session, tray — all 'cargo tree -i aws-lc-sys' => no match). Code already
used rustls::crypto::ring::default_provider (library.rs); no behaviour change.
Validated: cargo tree (all 4 windows crates aws-lc-free) + .21 clippy -D warnings
(pf-client-core, pf-presenter, punktfunk-client-session) clean.
The Windows host clippy CI broke building aws-lc-sys 0.41.0 (a heavy C dep that
cl.exe fails to compile on the runner). The host was the ONLY thing pulling it:
its rcgen used features=["aws_lc_rs"], and bare `rustls`/`tokio-rustls` take
the default (aws-lc-rs) backend, and gamestream/{mod,tls}.rs installed the
aws_lc_rs crypto provider explicitly — a pre-existing deviation from the
project's documented ring-only design (punktfunk-core + the client are ring).
Align the host to ring: rcgen/rustls/tokio-rustls pinned to the `ring` feature
(keeping tls12 for Moonlight/GameStream TLS-1.2 clients) and the two GameStream
provider installs switched to `rustls::crypto::ring::default_provider()`. Both
providers offer the same standard TLS 1.2/1.3 suites, so it's transparent to the
TLS paths; aws-lc-sys/aws-lc-rs leave the tree entirely (`cargo tree -i
aws-lc-sys` => no match). Not caused by the preceding log fix (that touched no
crypto deps / the lock).
Validated: cargo tree (aws-lc-sys gone, windows target, amf-qsv,qsv features) +
.21 host clippy -D warnings (--features nvenc) clean. Owed: GameStream on-glass
re-check with a real Moonlight client (low risk — same TLS suites, client is
already ring).
The W7/W8 client refactor (extracting FullscreenController/ApprovalRequest out of
ContentView, splitting Settings) landed the DefaultsKeys + scaffolding but dropped
the feature WIRING, which was uncommitted WIP shelved in a stash during the
16:28 clipboard/v4 reconcile. Recovered from that stash (+ its untracked-files
parent for the whole new ModifierLayout.swift) and re-ported onto the refactored
tree:
- invertScroll: applied in InputCapture.sendScroll (the one scroll sink) + Settings toggle
- modifierLayout (Cmd/Option switch): restored ModifierLayout.swift enum + KeyMaps
.applyModifierLayout + InputCapture.emitKey wire-boundary + Settings picker
- fullscreen shortcut (Ctrl+Cmd+F): InputCapture ⌃⌘F interception + onToggleFullscreen
+ StreamView wiring + Stream-menu item (the .punktfunkToggleFullscreen sink +
FullscreenController observer already survived on main)
- render scale: Settings picker + MatchWindowFollower renderScale/maxDimension
application + StreamView/StreamViewIOS plumbing (RenderScale.swift already on main)
StreamCommands re-ported by hand (the stash hunk deleted the since-landed clipboard
item — kept it, added the fullscreen item alongside). Recovered ModifierLayout +
RenderScale tests. swift build green; 15 tests pass.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The `wasapi` audio crate polls the default output device ~1×/s and
`log::debug!`s it (`wasapi::api default device Ok(...)`), flooding host.log
even at RUST_LOG=info: those `log`-crate records reach tracing through the
tracing-log bridge, and at *filter* time they carry the bridge's "log" shim
target, so neither a bare-info EnvFilter nor a `wasapi=warn` directive on the
file layer matches them (the ring already drops them post-hoc via
normalized_metadata + NOISY_DEBUG_TARGETS — the fmt layer filters pre-event and
can't).
Fix at the source: a shared `log_capture::install_global` replaces
SubscriberInitExt::init in both init paths (service file log + interactive
stderr) and inits the tracing-log bridge with `ignore_crate("wasapi")`, so the
poll records are dropped before they ever become tracing events — both sinks go
quiet, while every other log-crate dependency still reaches the console ring
(bridge max-level stays DEBUG). `wasapi` also added to NOISY_DEBUG_TARGETS as
belt-and-suspenders, and `log` moved dev-dep -> dep for LevelFilter.
Validated: .21 clippy (host --features nvenc) clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Ship the management web console and the plugin/script runner across the packaging
matrix, and fix a Linux tray startup crash.
- nix: add punktfunk-web (Nitro SSR on bun) and punktfunk-scripting (Effect runner),
with services.punktfunk.{web,scripting} modules. Both bun node_modules come from
fixed-output derivations pinned to real deps hashes; patchShebangs the web
node_modules so orval/paraglide/vite run in the sandbox. Validated on x86_64-linux
(nix build .#punktfunk-{web,scripting,tray,host} all green).
- punktfunk-scripting now ships for deb/rpm/arch/nix/windows, mirroring the web-console
mechanics (vendored bun on deb/rpm/arch, pkgs.bun on nix; one self-contained
`bun build --target=bun` bundle). Opt-in: host Recommends it, but the systemd --user
unit ships un-started (Windows registers its scheduled task disabled).
- fix(tray): build punktfunk-tray in its OWN cargo invocation on nix/rpm/arch. Cargo
unifies features across one `cargo build`, so co-building the tray with the host pulls
the host's ashpd->zbus/tokio onto the tray's shared zbus and it panics at startup
("there is no reactor running, must be called from the context of a Tokio 1.x
runtime"). Isolated, the tray's zbus stays on async-io. The .deb already did this;
runtime-verified on nix.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The iOS/tvOS/macOS-AppStore archive steps sign development-automatic OFFLINE
(no -allowProvisioningUpdates), so they depend on Xcode's cached managed dev
profiles. Adding the App Groups capability (group.io.unom.punktfunk, shared with
the Widget/Live-Activity extension) invalidated the cached "iOS Team Provisioning
Profile: io.unom.punktfunk" (it carries no application-groups entitlement) and
there is no managed dev profile for the embedded io.unom.punktfunk.widgets at all
— so the archive fails with "doesn't include the App Groups capability" and
"No profiles for io.unom.punktfunk.widgets".
Add -allowProvisioningUpdates + the ASC auth key (asc.p8 already written to
RUNNER_TEMP) to all three archive commands so Xcode syncs the App ID capabilities
and regenerates/creates the managed *development* profiles. This is development
signing against the Apple Development cert already in the keychain — no cert
creation, so the App-Manager ASC key suffices; the manual distribution export
steps are untouched.
Also drop the widgets IPHONEOS_DEPLOYMENT_TARGET from 27.0 (an iOS-27-beta-era
value, out of range on the runner's iPhoneOS26.5 SDK) to 17.0, matching the app.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
From the 2026-07-17 on-glass finding (.173 RTX 4090): the IDD-push capturer's
depth-2 pipeline measured ~13ms of glass-to-glass latency over depth-1 at
60fps (17ms → 4ms) — the AU is ready in µs but depth-2 holds it a whole frame
interval unpolled while N+1 is submitted. Depth-2 only earns its keep under
GPU contention (it overlaps the convert of N+1 with the encode of N, avoiding
the depth-1 ~50fps collapse). So the native encode loop now runs depth-1 by
default and escalates to the capturer's max ONLY when it can't hold cadence at
depth-1 (a leaky-bucket over 'the frame's work overran its pacing deadline',
with a startup warmup skip), then holds there for the session (no oscillation;
de-escalation is a v2 item). PUNKTFUNK_IDD_ADAPTIVE=0 pins the capturer's full
depth (pre-adaptive behaviour); no effect where the capturer's max is already
1 (every non-IDD backend). GameStream plane untouched.
Pairs with the shipped REALTIME auto-gate (1bc156aa): uncontended sessions get
the ~13ms latency cut, genuinely GPU-bound ones keep the depth-2 overlap AND
the REALTIME priority. Validated: .21 clippy -D warnings (host --features nvenc).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
design/latency-reduction-2026-07.md T2.5's Linux half: the LINEAR dmabuf path
(gamescope's only offer) fed NVENC RGB, paying its internal RGB->YUV CSC on
the SM the game is saturating — the exact contention §5.A removed everywhere
else. The Vulkan bridge now carries a buffer-to-buffer RGB->NV12 compute
shader (rgb2nv12_buf.comp, BT.709 limited, coefficient-identical to
pf-encode's rgb2yuv.comp; whole-word writes so no 8-bit-storage feature is
needed): import dmabuf -> dispatch CSC into the exportable buffer -> CUDA
de-strides both planes into a pooled two-plane NV12 buffer. PUNKTFUNK_NV12
(default-on) now covers LINEAR; a CSC failure latches RGB for the stream
(mid-frame fallback, no dropped frame); 4:4:4 LINEAR sessions stay RGB (never
silently subsample). New ImportKind::LinearNv12 rides the existing worker IPC
(appended last per the wire-tag rule); cursor stays downstream (blend_nv12).
Validated: .21 clippy -D warnings (pf-zerocopy/pf-capture/host+nvenc) + 17
zero-copy tests. Owed: on-glass gamescope session (visual + dmon sm% check).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Surfaced by giving windows-host CI its missing pf-encode/pf-frame path triggers:
the T2.3 auto-gate (dxgi.rs) and the clipboard host (windows.rs) landed with
Windows clippy owed — three undocumented unsafe blocks, one manual checked
division, one comment orphaned off its statement, and a latent E0432 in
standalone pf-clipboard builds (WNDCLASSW needs Win32_Graphics_Gdi; the host
graph only compiled via feature unification).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
--features nvenc,amf-qsv,qsv on build and clippy (static VPL dispatcher, no new
DLLs — ffmpeg *_qsv stays as the fallback until Phase 4). Adds pf-encode/** +
libvpl-sys/** to the paths trigger (drift from the W6 split: encoder changes
only reached this workflow via Cargo.lock luck). Exports
CMAKE_POLICY_VERSION_MINIMUM=3.5: audiopus_sys' vendored opus fails a FRESH
CMake-4 configure (observed on this runner on a clean target dir), so green
runs were one cache purge from breaking.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
design/latency-reduction-2026-07.md tier 2, the two code-side halves:
- T2.2: the Linux direct-NVENC backend gains the two-thread retrieve
(PUNKTFUNK_NVENC_ASYNC, the same opt-in knob as Windows): the session stays
sync-mode (async events are Windows-only) but the blocking lock_bitstream
moves to a dedicated pf-nvenc-out thread — the NVENC guide's sanctioned
submit-thread/output-thread split. poll() drains completions non-blocking,
submit() backpressures at PUNKTFUNK_NVENC_ASYNC_DEPTH (default 4) in-flight;
map/unmap and every other session call stay on the encode thread; teardown
joins the thread before destroying the session. Under a GPU-saturating game
completed frames queue instead of serializing capture on the encode wait.
- T2.3: PUNKTFUNK_GPU_PRIORITY_CLASS gains 'auto' AND IT IS THE NEW DEFAULT
(gpu-contention §5.C): HIGH immediately, then REALTIME where the documented
NVIDIA+HAGS+near-full-VRAM NVENC hang cannot bite — HAGS probed once via
D3DKMT WDDM_2_7_CAPS (off => REALTIME outright); HAGS on => a pf-gpu-prio
monitor flips REALTIME<->HIGH on LOCAL-segment VRAM headroom (downgrade
>92% of budget, restore <=85% for 3x2s polls). 'high' restores the old
static default; 'realtime' pins it (operator owns the hazard).
Validated: .21 clippy -D warnings (punktfunk-host --features nvenc) against
the QSV-merged main; .133 Windows cargo check of pf-frame + punktfunk-host.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
8/8 live tests green on the Arc Pro laptop (VPL 2.16): H264/HEVC SDR, HEVC
Main10+HDR metadata, AV1 10-bit, LTR-RFI clean recovery (no IDR; driver honors
mfxExtRefListCtrl on H264/HEVC/AV1), no-IDR bitrate retarget. Zero-copy
GetNativeHandle path confirmed (single textures, no array-slice hazard).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Reconciles the qsv line (libvpl-sys + qsv.rs + dispatch/probe wiring, verified
on the .133 Windows box: check + clippy + dispatcher link test + unit tests)
with the concurrent clipboard JPEG/GIF pass-through commits; Cargo.lock union
re-adds the libvpl-sys entry.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Transcoding every image to PNG was the Phase-1 floor, but it bloats lossy
originals (a copied JPEG re-encoded lossless is 5-10x the bytes for zero
quality gain — feeding the render-timeout on constrained links) and flattens
GIFs to one frame. The wire already carries kind LISTS, so offer the original
format beside the portable fallback and let the destination pick:
- New wire kinds image/jpeg + image/gif (§3.5 extension; strings only, no
protocol change). image/png stays the universal floor every peer accepts.
- macOS: public.jpeg / com.compuserve.gif rows pass through verbatim both
directions; the PNG floor is announced whenever any image is present.
- Windows: registered JFIF/GIF formats read + promised verbatim; every image
offer still promises CF_DIB, whose delayed render now fetches the richest
offered kind (png > jpeg > gif) through the generalized image_to_dib.
- Linux maps gain the matching rows (native MIME pass-through).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
design/latency-reduction-2026-07.md tier 1, remaining halves:
- T1.1: the native encode loop wakes on the capture's ACTUAL arrival instead of
sampling at a free-running tick — deletes the sample-and-hold (~half a frame
interval on average, a full one worst-case: ~8ms avg @60fps). New
Capturer::supports_arrival_wait/wait_arrival pair (IDD-push waits its
frame-ready event against the shared-header token; the PipeWire portal blocks
its channel with a pending stash); backends without an arrival signal — and
PUNKTFUNK_FRAME_DRIVEN=0 — keep the legacy tick bit-identically. A
0.9x-interval rate floor caps encode at ~1.11x target when the compositor
outruns the session; a +0.5x-interval keepalive keeps static desktops
re-encoding at 1.5x-interval cadence. Pacing deadlines re-anchor to the
actual submit so they can't drift against the arrival clock. GameStream
plane untouched.
- T1.4: the jump-to-live detectors run on WALL-CLOCK now (STANDING_TIME /
FLUSH_AFTER = 250ms) instead of 30-frame counts whose meaning scaled with
fps (500ms @60 but 125ms @240 — and stretching further under T1.1's slower
static-scene repeats). The queue trip also requires depth still >= high, so
a hysteresis-band hover can't fire on elapsed time alone.
Validated: .21 Linux 185 core + 177 host + pf-capture tests, clippy
-D warnings; .133 Windows cargo check of pf-capture + punktfunk-host green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
design/latency-reduction-2026-07.md T0.1/T0.2/T1.2/T1.3:
- T1.2 rate-capped front-loaded pacing: the paced overflow's budget is now
min(0.9x slack, overflow wire time at ~3x the live encoder bitrate)
(PUNKTFUNK_PACE_FACTOR, 0 = legacy deadline-only spread). A 300 KB-1 MB
frame's tail leaves in ~2-5 ms instead of smearing across ~15 ms at 60 fps;
GameStream schedule byte-identical (pins unchanged).
- T1.3 data-first wire order: packetize emits every block's data shards before
any parity (per-block parity pools keep all blocks' parity alive for the
second pass), so lossless completion stops waiting behind the parity tail.
EOF = last emitted packet; receiver already order-agnostic.
- T0.1 staged 0xCF: HostTiming gains an append-extensible per-stage tail
(queue/encode/pace us; seal+channel-wait derived as residual) - no cap bit
needed, old peers read the 13-byte prefix. Joined client-side into
Stats::host_{queue,encode,xfer,pace}_ms, the OSD detailed tier, and the
probe's report.
- T0.2 true on-glass present timing: VK_KHR_present_id/present_wait enabled
when supported; a PresentTimer waiter thread resolves each present id to
real visibility, replacing the submit-time display stamp (which undercounts
by up to a refresh and hides a silent-FIFO standing queue).
Validated on .21: core 185 + host 185 tests, pf-presenter 19, clippy
-D warnings across all five touched crates; loss-harness recovery curve
unchanged; C ABI harness round-trips.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
macOS image copies rarely carry public.png — screenshots/Preview put TIFF on
the pasteboard, browsers add WebP/AVIF/GIF (observed live: TIFF+RTFD+WebP+AVIF+
8BPS+GIF, no PNG) — so the literal .png announce never fired and images
silently didn't sync. Announce image/png whenever a convertible image is
present (TIFF/HEIC alongside the native PNG check) and convert at serve time
via NSImage -> NSBitmapImageRep PNG (lazy, per design §3.5 — bytes still cross
only on a host paste).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
macOS image copies rarely carry public.png — screenshots/Preview put TIFF on
the pasteboard, browsers add WebP/AVIF/GIF (observed live: TIFF+RTFD+WebP+AVIF+
8BPS+GIF, no PNG) — so the literal .png announce never fired and images
silently didn't sync. Announce image/png whenever a convertible image is
present (TIFF/HEIC alongside the native PNG check) and convert at serve time
via NSImage -> NSBitmapImageRep PNG (lazy, per design §3.5 — bytes still cross
only on a host paste).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Brings origin's release prep (widgets App Store profile mapping + workspace
version 0.13.0) into the local line that landed the Linux stream sink, the
Windows client-only audio rework, the July security-audit fixes, cross-client
render-scale, and the widgets scheme. release.yml was byte-identical on both
sides; no other file overlap.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Mirrors origin's 15233a68 (byte-identical release.yml, so it dedupes at the
next reconcile) and adds the shared PunktfunkWidgetsExtension scheme the
archive step builds with.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Client-side supersampling/downscaling: the client asks the host to render
and encode at chosen-resolution × scale (the host does no scaling) and the
presenter rescales the decoded frame to the display. >1 supersamples for
sharpness; <1 lightens the host GPU and the link. Default 1.0 = Native, the
prior behavior.
The geometry lives once in punktfunk_core::render_scale (multiply, preserve
aspect ratio, floor to even, clamp to the codec's per-axis ceiling — 4096
for H.264, 8192 otherwise), the Rust twin of the Apple client's
RenderScale.swift, consumed by the native session client, the presenter's
match-window path, the Windows/Linux settings UIs, Decky, and Android
(settings + host connect + unit test).
Implemented and platform-verified by the Apple-client-features session
(Linux+Android+Apple green there); the punktfunk-core wiring
(pub mod render_scale) is restored here after being lost in a working-tree
reconciliation.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The low/medium findings from the July host+Windows security review, as
implemented in the audit session's working tree:
- Webhooks and library-art fetches refuse loopback/link-local/metadata
targets and no longer follow redirects (SSRF pivots from the privileged
host process).
- The paired-client rosters (/clients, /native/clients) move off the
streaming-client auth lane — one paired device could enumerate every other
device's name + fingerprint; only the bearer/loopback console keeps them.
- Device-name sanitizing extends to bidi/format control characters
(spoofable rendering) via the shared native_pairing::is_spoofy_char;
stream-marker quoting uses the same set.
- The SYSTEM service resolves powershell by its full System32 path —
CreateProcess checks the launching EXE's own directory first, so a planted
powershell.exe beside the host binary would have run as SYSTEM.
- The pf-vdisplay driver's opt-in file log moves from world-writable
C:\Users\Public to WUDFHost's own temp dir.
- GameStream pairing sessions are single-use (removed whatever the outcome).
- Uninstall also removes the pf_mouse driver-store entry (rider from the
virtual-HID-mouse work).
- openapi.json regenerated (hardened-config-dir doc wording).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Root cause of the field report (Android client, Windows host: no audio until
the user manually cycled Sound-output devices, then audio on BOTH PC and
phone): the WASAPI loopback captured whatever the default render endpoint
was at open time — not the wiring plan's chosen endpoint — the plan's
IPolicyConfig default-set is warn-only and racy, and nothing reacted to
mid-stream default-device changes.
- Explicit-endpoint capture: the capture thread opens the plan's
loopback_render by id, never "the default" (KEEP_DEFAULT preserves the old
default-capturing behavior with the echo guard).
- Client-only playback default: wiring_plan::plan(..., host_audio) prefers a
silent sink (Steam Streaming Microphone render side — loopback-validated,
silent on host) over real hardware, so stream audio plays on the client
only; PUNKTFUNK_HOST_AUDIO=1 restores real-hw-first (audible on the host).
The capture side auto-installs the Steam pair once per process when no
silent sink exists; open() handshake timeout 3s -> 30s to cover it, and a
handshake timeout now stops the detached thread (it used to run for the
process lifetime with the default still parked).
- Self-healing capture thread (wasapi_cap): outer capture_once loop
(Assert|Follow) with a ~1s watchdog on the default render id. A user
switch to a capturable endpoint is followed (their choice wins, audio on
both); a switch to a dud (cable/SSS/mic target) re-asserts the plan;
IPolicyConfig-denied converges to Follow instead of churning. Device
errors reopen with 2s backoff; only the FIRST open failure is fatal.
Zero-packets breadcrumb after 30s distinguishes broken-loopback from
quiet-desktop.
- Park/restore of the default playback device (audio_control):
wire_now(set_playback) parks the default on the loopback sink only for the
capture's lifetime (the mic pump passes false — it runs while the host is
idle); crash marker audio-default.prev + recover_orphaned_default() at
first wire; restore is skipped if the operator changed the default
themselves. A mic-default hygiene pass keeps VB-Cable installs audible and
never records the mic target as the restore target.
- Session-end park_audio_capture(): Windows DROPS the capturer (thread join
restores the default) instead of caching it; Linux keeps the parked
PipeWire thread. Composes with the stream-sink idle() hook at all three
park sites (idle is a no-op on Windows).
Verified: Linux (.21) clippy -D warnings + 176 punktfunk-host tests green
(incl. the new wiring-plan preference tests); Windows (.173) clippy with
nvenc,amf-qsv --all-targets -D warnings green at this exact tree.
On-glass winbox/Android validation still owed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Merges 44b71e74 (see its message for the full design: the "Punktfunk Stream
Speaker" Audio/Sink claimed as the session's default output — the crackle
root fix for hardware-sink churn — plus the core-error liveness fix and true
5.1/7.1 capture) across the W1–W8 refactor that landed since the branch
forked:
- punktfunk1.rs was split on main: the audio_thread idle()/drain() hunks are
hand-ported into native/audio.rs (same three sites: reuse-drain comment,
encoder-fail park, end-of-thread park).
- pwinit moved to pf-capture (W6.2): the branch's crate::pwinit calls in
pw_thread and stream_sink.rs become pf_capture::pwinit.
- Kept main's log-tier demotions (stream-state + setup lines at debug)
inside the restructured pw_thread.
Resolution verified on Linux (home-worker-5): clippy -D warnings clean +
173 punktfunk-host tests green at the 691c064a-based equivalent ba5973a2;
re-verified at this base before landing.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The eight W6 leaf crates hardcoded 0.12.0 instead of inheriting the
workspace version — switched to version.workspace = true so the next bump
is one line again.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
