Commit Graph
5 Commits
Author SHA1 Message Date
Mike-Solar adf2cef32c feat(oakrender): render-process isolation S3 - audio over shm, per-ticket slot formats, tuning
- Audio tickets join the process backend: render_audio_batch wire
  message, workers mix straight into shm slots (SLOT_FORMAT_AUDIO_F32),
  ShmAudio payload with release semantics, crash isolation covers audio
  renders; playback audio uses an async 4-chunk prefetch drained on the
  UI tick (also fixes the sub-60fps chunk truncation bug); oversized
  ranges and dispatcher outages fall back to in-process inline.
- Per-ticket slot formats: force_format is honored (exports request
  F32 slots, dropping the BGRA8 round-trip and its 8-bit quantization);
  segments grow on demand via worker-idle rebuild with generation
  handoff; the scheduler filters over-capacity tickets.
- Adaptive defaults: 128-256MB/worker segment budgets drive slots per
  worker, batch size follows workers/slots; bench_process example
  measures throughput and adjacent-frame completion deltas
  (e.g. 4 workers: 841 fps, 4.6ms mean delta).
2026-08-19 00:42:03 +08:00
Mike-Solar cad1d93544 feat(oakrender): render-process isolation S2 - process backend by default, zero-copy onscreen
- WorkerPool thread pool deleted; RenderManager defaults to the
  Processes backend (oak-worker children), Threads kept as a test-only
  inline dispatcher; audio tickets stay in-process until S3.
- Onscreen path reads worker shm slots directly: BGRA8 slot format,
  RenderedFrame::Shm wrapped into the display buffer (single disclosed
  GPU-staging memcpy), scopes analyze BGRA8; the long-lived full-res /
  thumbnail paths take the counted slot_to_vec copy and release.
- Playback pre-render window: forward 120 frames (configurable) fed to
  the PreviewScheduler at Playback priority, interleaved across
  workers, cached in shm slots until the playhead consumes them;
  generation-based invalidation cancels and releases on edits.
- oaktask export and oak-cli run on private ProcessDispatchers (fixed
  a pump-while-locked self-deadlock in the export loop); facade
  get_frame handles ShmFrame payloads.
- Acceptance: preview path main_heap_frame_copies == 0 with spawned
  workers, CLI transcode/render verified end to end.
2026-08-18 20:45:24 +08:00
Mike-Solar 431b9ed2b1 feat(oakrender): render-process isolation S1 - dispatcher, scheduler, real worker
Per the M15 design (docs/zh/plans/riir/M15-render-process-isolation.md):

- ipc.rs moved into oakrender with protocol v2: hello_caps,
  render_batch, batch_accepted, frame_failed; main-process-assigned
  slots; BGRA8 slot format. POSIX shm verified to 1GiB on macOS.
- ProcessDispatcher: spawns oak-worker processes, handshake, stdio
  NDJSON control, shm segment lifecycle with generation-tagged keys,
  crash detection with bounded restart and frame redispatch, zero-copy
  ShmFrameRef delivery and copy counters.
- PreviewScheduler: interleaved batch claiming (frame % W per worker,
  no work stealing), seek > playback-distance > background priority,
  credit-based flow control, crash recovery.
- oak-worker renders for real: graph snapshot deserialization, montage
  decode+composite straight into the assigned shm slot, F32->BGRA8
  final conversion in-worker, OFX plugin executor installed in-worker,
  crash hooks for isolation testing.

Thread pool coexists for now (S2 removes it). Integration tests cover
two-worker zero-copy rendering, crash isolation with redelivery, and
real H.264 footage decode into slots.
2026-08-18 18:58:38 +08:00
Mike-Solar b36cbd6b6f refactor: purge CHandle from module internals (M14 R5)
Module-internal object references are Rust types now (values, Arc,
Mutex); CHandle remains only at the oakengine C-ABI boundary:

- oakundo: the global stack holds UndoStack/UndoCommand values
  directly (stack token is the static's address)
- oaktimeline: marker/workarea boxes carry Arc<Mutex<T>>; commands
  share the same allocation through Arc clones (readers in oakengine
  stubs and the app's graphops updated to lock)
- oaktask/oakstorage: sessions, write-through bindings and the
  database backend pass ProjectArc; the Session drops its manual
  release bookkeeping; nodeutil keeps the CHandle<->Arc boundary
  conversion (release_project restored for the app)
- oakcodec: handle.rs deleted outright (no facade entry needed it);
  texture/block placeholders are unit structs
- oakrender: copier's project handle is an identity u64; alive-count
  machinery removed; handle.rs is make_owned/get/get_mut only
- oakplugin: the instance registry is gone (its unregister key never
  matched, leaking weak entries); handle.rs is the RefBox boundary type
- oaknode/oakcommon: only dead guard/borrow helpers removed; external
  payload handles (texture/processor) documented as the boundary

Flake hunts landed along the way: the audio recording test serializes
on the shared manager lock with a normalized state; the autocacher
cancel test uses a slow producer so cancellation is deterministic.
2026-08-17 16:40:15 +08:00
Mike-Solar 013a175707 refactor: workspace layout — crates/, app at root, legacy C++ removed
Single mechanical restructure commit:
- root Cargo.toml = oakapp bin + workspace; one cargo build produces
  oakapp, oak-cli, oak-worker, liboakengine.dylib
- app/rust/src -> src/ (app at repo root, no rust/ nesting)
- src/<mod>/rust -> crates/oak<mod>; src/oakcore-rs -> crates/oakcore;
  src/bindings/oakotio -> crates/oakotio; src/engine/rust ->
  crates/oakengine (keeps cdylib+staticlib+rlib)
- public C headers include/<mod>/ -> crates/oakengine/include/<mod>/
- OFX SDK headers vendored into crates/oakplugin/ofx/ (HostSupport gone)
- legacy deleted: old src/ C++ modules, engine/, core/, ffmpeg_bridge/,
  app/ (Qt), cli/worker C++, root CMakeLists, third_party/KDDockWidgets
  submodule, otio-install, all build-* output (~40GB)
- oakstorage kept but excluded from the workspace (skeleton w/ todos);
  gpui excluded (own workspace)
- verified: cargo build green, cargo test --workspace 1845/0
  (with the documented OCIO_RS_* env override for the homebrew OCIO)
2026-08-10 20:24:25 +08:00