- oakrender: new ticket family (render_frame/render_audio with finished
callback, result frame/samples access, cancel/wait), project copier
C API, cache get_invalidated_ranges, manager set_aggressive_gc
- oaknode: node_copy_inputs, set_value_hint_track, viewer params
setters, video frame cache borrowed outlet, project copy_settings
- oakcodec: encoder desired pixel format, export format extension,
encoding generate_matrix, custom range in encoding params POD
- oaktask: RenderTask orchestrates oakrender tickets (no Qt threads/
watchers), PreCacheTask and ExportTask rewritten over the oaknode/
oakrender/oakcodec C ABIs; frames cross from oakrender to oakcodec
handles by pixel copy (different control blocks)
- the two-step texture/download render path collapses into single-step
(tickets always yield frames); subtitle sidecar kept
- tests: 105 in build-oaktask (export/precache factory paths,
construction, conform end-to-end); regressions all green
- Task base: Qt signals become lifecycle listeners (the async-command
callback exception), cancellation uses oakrender's OakCancelAtom C
handle; TaskManager runs std::thread workers, no signals
- ConformTask/ProxyTask implement oakcodec's submit-callback contract
(synchronous interim); register_codec_task_submitter() closes the
conform/proxy loop - conform of demo.mp4 produces pcm caches in tests
- ProjectImportTask/ProjectLoadTask/ProjectSaveTask over the oaknode
C ABI; image-sequence confirmation becomes a facade callback
(default: not a sequence)
- C API additions needed by oaktask: oaknode serializer file-level
save/load (auto-initializing), footage video-params/cancel-atom/
as-node, folder add-child command factory; oakcodec decoder
conform_audio + image-sequence helpers; oakrender cancelatom
get_native
- fix double-ownership: submitting a task to the manager transfers
ownership, freeing its handle only releases the wrapper
- C ABI in include/task (task/manager/project, OakTaskTask opaque
handle + lifecycle subscribe), every function tested (103 cases in
build-oaktask incl. regressions)
- RenderTask family and OTIO tasks follow in separate commits
- 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
- config moves into oakcommon as ConfigStore + oakcommon_config_* C
API (INI storage, typed entries, error-handler injection); node and
render call sites keep OAK_CONFIG() macro shape via a local shim
that forwards to the C API; transition config stubs removed
- oakaudio: de-Qt all six classes, C ABI in include/audio with
refcounted handles (processor/manager/waveform/levelmeter/sync,
48 functions); PreviewAudioDevice moved in from render; recording
goes through oakcodec encoder; waveform extract uses probe +
ffmpeg_bridge decode (decode_audio needs M8 task system)
- fix re_sum_samples min/max init bug (values clamped to 0 for
same-sign ranges)
- every C API function has positive + error-path tests; suites:
oakcommon 193, oakaudio 36, oaknode 96, oakrender 42, oakcodec 18
- oakcodec: de-Qt all 20 sources (QThread decode loop -> std::thread,
QObject/signals -> callbacks), pure C ABI in include/codec with
refcounted neutral handles (OakFrame/OakDecoder/OakEncoder),
framemanager moved in from render, frame_to_buffer/buffer_to_frame
moved in from oakcommon oiioutils, codec->task via submit callback
(M8 will register), all cross-module calls go through the other
side's C API, -fvisibility=hidden + OAKCODEC_API
- oakcommon: handles become refcounted value structs
{ctx, addref, release, abi_version} (FFmpeg-style), pass-by-value
signatures, free() as release wrapper; init_from_native/get_native
for copyable value objects; OakCommonXxx renamed to OakXxx
- oakcommon: add logging (log_debug/info/warning/critical with level
filtering and sink injection) + printf-style oakcommon_log C wrapper
- oakrender: add CancelAtom C API family; complete
oakrender_color_processor_convert_frame; fix get_processor() missing
definition and OCIO env var lookup
- tests: oakcommon 174, oaknode 96, oakrender 42, oakcodec 18, all
green in their standalone builds
- 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
Mike-Solar