- 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.
The static FFmpeg's external codec libs pull in -lz, which on this
toolchain resolves to a copy whose install name is @rpath/libz.1.dylib
(zlib-ng-compat); without an LC_RPATH entry all three binaries died in
dyld at startup. The app/cli/worker build scripts now emit
-Wl,-rpath,/usr/lib.
Also: FFMPEG_DIR moves into the committed .cargo/config.toml as a
workspace-relative [env] entry — ffmpeg-sys-next's build script cannot
read .env files, and without it the crate silently linked the shared
Homebrew FFmpeg while oakffmpeg-link emitted the static transitive
flags (mixed linkage). docs/build.md updated.
- oakengine is now cdylib-only (no rlib/staticlib consumers anywhere;
cargo tree verified) — the plugin/external C ABI layer; README and
docs updated
- cd.yml drops the dylib embedding/re-sign steps (the app no longer
links it)
- test race root-caused and fixed for good: the global undo stack lock
is now a re-entrant mutex (parking_lot) shared by every test that
drives the stack, including the previously unlocked node/render
families; the render-manager serial-ordering bug (an earlier repro
test initialized the global manager before the not-initialized test)
is fixed with a shared SERIAL guard and a manager shutdown
- 5 consecutive parallel runs clean; serial 209/209
- viewers show the 480px proxy immediately and a background thread
fills the sequence-resolution frame (per-monitor in-flight job,
generation-based staleness, playback skips full-res)
- preferences dialog complete: cache dir (now consumed by
default_disk_cache_path), proxy policy/divider, snapshot interval
(write-through era autosave), default transition length, audio
in/out devices (new facade device-enumeration exports; audio init
from config — playback was never creating the audio instance),
language/theme/renderer backend, all persisted via config
- shortcut map (src/shortcuts.rs): space/J/K/L, I/O, S split, A/^A,
⌘Z/⌘⇧Z, ⌘N/⌘O/⌘S/⌘E/⌘Q, frame step, Home, track zoom; dispatch
shares the menu action path and stays silent over modals
- screenshots: preferences dialog zh/en captured and reviewed
D3: oakdb+pg:// fully wired (shared sea-orm entities, BIGSERIAL DDL,
connect-probe instead of pool retry on dead servers); Storage/Backend=pg
+ Storage/PgUrl config; 13 OAK_TEST_PG_URL-gated PG tests (verified
against a Docker postgres:16), always-on clean-error tests otherwise.
D4: DaVinci-style project manager — list with derived stats, create/
rename/duplicate/delete (confirm)/import/export (native dialogs,
ove/otio/fcpxml), shown at startup and from the file menu; facade
oakengine_library_* exports (list/create/delete/rename/duplicate/
import/export + project_load_library that binds write-through);
save/save-as menu becomes 'export project file', open splits into
from-library/from-file; status bar shows library write state; storage
activates on app start and flushes on exit; spawn_modal reentrancy
fixed (window-callback path) with a doc note.
Also: the P1 audio test's environment probe was lost in the ffi purge;
restored on cpal (the output device is cpal now).
- facade storage session manager: project handles bind to the default
SQLite library on project_new/load; every undo push/group_end/jump
write-throughs via DatabaseBackend::save (diff journal); project_free
flushes and unbinds
- background snapshot thread (Storage/SnapshotIntervalSec, latest-wins,
newest 3 kept) with exit flush (oakengine_storage_flush)
- config-gated: storage only activates with an explicit
Storage/Backend=sqlite, so headless consumers and tests never touch
the real library; new exports: storage_flush/is_bound/last_error
- it_storage: kill -9 recovery, cross-session undo, snapshot pruning,
multi-project isolation, graceful degradation
- also fixes a real config test polluting the user config.ini and the
undo-stack test races
- oakcore-rs: Rational/TimeRange/TimeRangeList/PixelFormat/SampleFormat
mirroring oakcore C++ semantics (91.6% coverage)
- oakundo: UndoCommand/UndoStack with vtable commands (99.6%)
- oakcommon: config/XML/VideoParams/logging etc. (89.8%, 516 tests);
XML via quick-xml, logging via the log facade
- oaktimeline: markers/workarea/edit commands (95.4%)
- oakaudio: processor/manager/sync/waveform/levelmeter (88.4%)
- unified -MMCCCC error codes across all C ABI headers (registry in
include/common/error.h; pass-through rule)
- olive::Variant moved node -> common (cross-module value type)
- oakcommon_videoparams_set_is_3d added to the C ABI
- M11 (oakplugin Rust rewrite) plan + notes.md freeze line for the
inter-module C ABI wiring
- oakrender cache C API: add create_for_node (four cache kinds),
get_uuid, request, load/save_state, set_saving_enabled,
set_passthrough, get_passthroughs, get_valid_cache_filename,
get_timebase, lock/unlock, invalidate_range (rational variant),
get_native (C++ only, for the in-module PreviewAutoCacher)
- Node's four caches become owned OakRenderCache value handles; all
cache access in node.cpp/clip/viewer/traverser/serializers goes
through the C ABI
- oaknode_node_get_video_frame_cache returns an addref'd
OakRenderCache; drop OakNodeFrameCache; oaktask precache/export and
oakrender_video_ticket_params take the value handle
- color: OCIOBaseNode/OCIOLutNode hold OakColorProcessor handles
(dtors added); oakrender gains color_processor_create_transform/
create_lut/create_grading_primary/get_native and LUT library
queries, keeping all OCIO transform construction inside oakrender;
oaknode gains colormanager_wrap_borrowed/get_native
- RenderManager::instance() check -> oakrender_manager_available();
cancel_video_tasks and disk cache path go through manager C API
- remaining node->render C++ symbol: only olive::Texture::~Texture
via Variant's shared_ptr payload (recorded exception, same class as
the UndoCommand inheritance)
Tests: new cache/color/manager/colormanager cases; all standalone
trees green (common 196, codec 22, audio 40, task 112, render 63,
timeline 125, node 98, plugin 102).
- ViewerOutput owns timeline markers/work area via refcounted
OakTimeline* value handles (oaktimeline_*_create) instead of C++
objects
- oaktimeline C API: add marker_list_create/marker_add/
workarea_create/workarea_set_enabled; handle boxes gain an optional
deleter for owning handles
- oaknode_node_get_markers/get_work_area return addref'd OakTimeline*
handles via out params (named-struct forward declarations avoid the
public header cycle); drop OakNodeMarkerList/OakNodeWorkArea
- serializers 210528-230220 load/save markers and work areas through
the oaktimeline C API; SerializedMarker POD replaces
std::vector<TimelineMarker*> in the serializer interface
- Sequence::add_default_nodes uses oaktimeline_add_track_command
- clip.cpp display-mode enums come from the new neutral
include/timeline/displaymode.h; delete src/node/transition/timeline
- oakcommon xml: add oakcommon_xml_reader/writer_wrap_native borrowed
wrappers (C++ adapter use)
- liboaknode now has zero C++ symbol references into liboaktimeline
Tests: new owning-handle and wrap_native cases; all standalone trees
green (common 196, codec 22, audio 40, task 110, render 49,
timeline 123, node 96, plugin 100).
- de-Qt src/plugin/src (olivehost/oliveplugininstance/oliveclip/
paraminstance/image/pluginprogressreporter); QMessageBox/
QApplication replaced by facade callbacks
- new C ABI in include/plugin/{error,host,instance}.h with
refcounted OakPluginInstance value handle
- paraminstance bridges via oaknode C ABI (OakNodeNode value handle,
oaknode_node_identity registry); undo via oakundo C ABI
- oliveclip textures are OakRenderTexture value handles; oakrender
gains texture/copier/ticket C API additions
- oaknode gains get_input_at_time/set_input_at_time_undoable/
node_identity/sequence_from_node/sequence_set_default_parameters/
find_input_footage
- move avframeptr.h to src/common/src (shared by codec and render)
- node transition pluginSupport stubs bridge the real oakplugin
headers; all oaknode-building standalone trees add src/plugin and
link oakplugin into their test binaries
- plugin tests self-sufficient (ipc shim, OfxHost force_load,
PRE_TEST discovery, OCIO env); timeline standalone gains
render/c_api
- restore ffmpegdecoder.cpp in src/codec/src/ffmpeg/CMakeLists.txt
(dropped in 3d004c081, caused jump-to-0 in decoder/encoder tests)
All six standalone trees green: codec 22, audio 40, task 110,
render 49, timeline 121, plugin 100.
- de-Qt all 11 sources: markers/workarea lose QObject/signals (list
re-sort via direct resort() call, drawing moves to the app layer),
explicit unique_ptr ownership for marker lists and viewer's
workarea/markers
- timeline undo command families (split/pointer/ripple/general/track)
now hold oaknode C handles and do all graph work through the oaknode
C ABI (no adapter layer); gaps are attached to the project graph
before insertion, orphan handles tracked and freed
- fix the de-Qt severed Block->Track length-changed chain with a
direct block_length_changed() call (positions were stale after
split/trim)
- expand oaknode C API by ~20 functions needed by timeline
(copy_in_graph, block kind/casts, connect_element, input arrays,
tracklist family, marker/workarea borrowed outlets)
- pure C ABI in include/timeline (marker/workarea/edit, command
factories returning OakUndoCommand), oakcommon XML get_native
accessors
- oaknode<->oaktimeline resolve each other at runtime; standalone
drivers link both into test binaries
- tests: oaktimeline 117, regressions oakcommon 193 / oaknode 96 /
oakrender 42 / oakcodec 18 / oakaudio 36 all green
- copy engine/render to src/render/src (sunk param types excluded),
de-Qt in five parallel groups: core machinery (tickets/worker pool/
jobs), caches, color/texture, preview/IPC, GPU backends
- replace Qt GL/Vulkan wrappers with native context abstractions
(CGL/EGL/WGL, raw vulkan.h), QProcess with POSIX WorkerProcess,
QJsonObject with a minimal NDJSON-compatible workerjson (wire
protocol unchanged), QDataStream disk state with a byte-compatible
BinaryStream
- signals become single std::function callbacks or facade-triggered
calls per the documented signal/slot strategy
- pure C ABI in include/render + src/render/c_api (renderer/cache/
color/manager families, OAKRENDER_E_* codes), 37 gtest cases
- bridge src/node/transition/render/* stubs to the real oakrender
headers, closing the node<->render cycle: liboaknode links
liboakrender, zero dangling symbols
- docs: M7 implementation status + oakrender semantic-change notes
- strip QAbstractItemModel/QAction/signals-slots from UndoStack (UI
concerns belong to the app layer), index_changed becomes a
std::function callback
- decouple UndoCommand from Project via an optional modified-flag
callback pair
- add pure C ABI in include/undo + src/undo/c_api (oakundo_ prefix,
vtable-based command wrapper, OAKUNDO_E_* error codes)
- fix three engine-side defects: jump(0) infinite loop,
MultiUndoCommand child leak, push_pre_executed not undoable
- add gtest suites (22 cases) and a standalone build driver
- de-Qt all classes under src/common (std::string/vector/mutex,
std::filesystem, expat-based XmlStreamReader/Writer)
- add pure C ABI in include/common + src/common/c_api: opaque handles,
init returns NULL on failure, free(NULL) is a no-op, out-params with
negative OAKCOMMON_E_* error codes (include/common/error.h)
- remove single-consumer classes from oakcommon (html, jobtime,
otioutils, playbackaudioclock, tohex, util, avframeptr,
crashpadinterface/crashpadutils, autoscroll, digit, range); their
destinations are recorded in docs/zh/plans/riir/notes.md
- add gtest suites under src/common/tests (127 cases), standalone
build driver in src/common/standalone
- riir.md: split project file read/write into standalone oakstorage
module (pluggable backend, DB-replaceable); add Mermaid data flow
diagram of all modules; adjust batch table and milestones
- riir/04-interfaces.md: new provides/consumes contract matrix for
all modules
- riir/M10-oakstorage.md: new manual with frozen C API (typed opaque
handles, manual vtable for storage backends, URI addressing)
- riir/01: codify pure-C OO interface rules (typed handles instead of
void*, init/free pairs, manual vtables, no C++ objects or member
calls across shared library boundaries); upper layers issue commands
only, no module-to-module event subscriptions (async task callbacks
excepted)
- sync 00/02/03 and M2-M9 manuals: remove subscribe APIs, replace
void* with typed handles, update tests accordingly
New app/playback/playbackcontroller.{h,cpp}: a lazily created
process-wide singleton whose set_playhead() forwards to
oakengine_viewer_set_playhead and re-broadcasts playhead_changed as a
plain Qt signal. Every app-side oakengine_viewer_set_playhead call site
(17 across viewer, timelinewidget, timelineview, import tool,
timebasedwidget, keyframecontrol, export dialog, mainwindow) now goes
through it.
Migrate the first four playhead subscribers off the raw C event
subscription / EngineEventBridge to the controller signal:
- TimeBasedView (issue 1)
- NodeParamViewWidgetBridge (issue 2)
- NodeParamViewKeyframeControl (issue 3)
- NodeParamViewConnectedLabel (issue 4)
Each stores a QMetaObject::Connection filtered on its viewer node and
disconnects on teardown; the viewer_sub_ members are gone. Plan docs
mark issues 0c and 1-4 as done.
Closes#30, closes#31, closes#32, closes#51, closes#52.
- app/ no longer includes engine C++ headers nor holds engine C++ types:
engine access goes through the oakengine C ABI plus C++ wrappers
(oakutil/oaknode.h, oakutil/oakvideo.h) and app-local mirror types
(tooltypes, trackreferencehandle, timelinecommonapp, keyframetypes,
subtitleapp, serializedlayoutinfoapp, nodevaluehandle, sliderdisplaytypeapp)
- engine: new C ABI functions for block/track/clip/transition navigation
and predicates, links, caches, waveform/playback, disk folder,
sequence_track_list, node_free, footage_is_valid, block_get_track,
get_brush; loadotio/saveotio ported to the current engine API
- OTIO is now a required dependency: CI and CD build it on every
platform, FindOpenTimelineIO fixed for OTIO 0.16/0.19 (the old deps
include requirement silently disabled OTIO everywhere), runtime
libraries are bundled into packages and copied next to macOS binaries
(oak_copy_otio_runtime)
- fix ProjectViewModel drag&drop mime read/write size mismatch (segfault)
- unify color label naming (k_olive -> "Oak") in the app-side mirror
- docs: OTIO required, FFmpeg minimum corrected to 6.0 (en/zh)
- gtest suite: 1925 passed, 0 failed
Move the pure-header utilities shared by app/ and engine/ out of
engine/common/ into a new shared/include/oakutil/ layer (define, lerp,
decibel, digit, range, crashpadutils, autoscroll, qtutils, filefunctions
declarations, and a trimmed xmlutils exposing a CancelAtom-free void*
overload). engine/common/ keeps forwarding headers so internal include
paths are unchanged; app/ now includes oakutil/* directly.
engine/node/project.h gains an explicit NodeGroup forward declaration
previously obtained transitively through the old xmlutils.h.
Move the finished migration campaign docs (handoffs v3-v6, roadmap,
R5 guides, R6 cleanup, R7 pure-ABI) from docs/zh/ into
docs/zh/plans/completed/ with an archive README; fix all
cross-references; refresh plans/README.md index (active plans now
marked unlocked).
Complete documentation set for the facade migration (B1-R6) and beyond:
facade-migration-roadmap, handoffs v3-v6, R5 guides, R6 cleanup plan,
R7 pure-ABI plan, long-term plans (riir/ai-agent/gtest/ui-redesign),
updated Chinese README draft, UI design mockups, and the Google Test +
struct-typedef rules in CONTRIBUTING.
Mike-Solar