Repository navigation
Conversation
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 1b2cfc57-c6c7-44ce-95c9-72368d409ca8
Add explicit recording status and errors, detect short/failed SD writes without blocking playback, and finalize WAV headers through the existing scheduler before closing files.\n\nCo-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Keep live recorder setup on the audio task and retry finalization queue pressure before reporting an invalid file.\n\nCo-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Drive initial WAV seek and header writes from the service state machine without blocking mixer output, and make SD scheduler enqueue genuinely nonblocking. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Preserve the scheduler contract used by existing decoders and encoders while keeping OutputWAV queue retries nonblocking. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
djdefi
added a commit
to djdefi/JayD-Library
that referenced
this pull request
Aug 25, 2026
PR CircuitMess#20 advanced to ac6ba80 ("Scope nonblocking SD enqueue to recording"), reverting SDScheduler::addJob() back to the legacy blocking void contract and adding a new nonblocking bool tryAddJob() scoped exclusively to OutputWAV's real-time recording write/finalize path. That makes the earlier 78a9f3d fix (which retrofitted bool-return checking onto every addJob() caller, matching an intermediate nonblocking-everywhere revision of CircuitMess#20) obsolete and non-compiling against the restored void signature. Revert SourceAAC::addReadJob/seekSourceFrame, SourceMP3::addReadJob, SourceWAV::addReadJob, and OutputAAC::addWriteJob to plain blocking Sched.addJob(...) calls with no return-value check, matching the restored contract. OutputWAV's two call sites already correctly use the new Sched.tryAddJob(...) (renamed by ac6ba80 from its prior addJob() bool usage) with the return value checked; no changes needed there. Queue-ownership review: Sched is drained both by the main sketch loop (LoopManager::addListener(&Sched)) and, at specific synchronous wait points, directly by the calling thread (e.g. MixSystem open/openChannel busy-wait via Sched.loop(0) while awaiting isReadReady()). The audio task's blocking addJob() calls (SourceAAC/MP3/WAV read jobs, OutputAAC write jobs) are drained by a different, independent thread/task, so a full queue blocks that call only until the drain thread services it; no self-deadlock. OutputWAV's write/finalize path runs on the same real-time audio task and must not block it, which is exactly why tryAddJob()'s nonblocking, checked, retry-on-false contract remains scoped there. Rewrite tests/JobQueueContractSelfCheck.cpp for the new dual-API contract: structural checks assert SourceAAC/SourceMP3/SourceWAV/ OutputAAC call only the blocking Sched.addJob and never Sched.tryAddJob, OutputWAV calls only Sched.tryAddJob with every call site's return checked and never the blocking Sched.addJob, and SDScheduler.h declares both signatures. A functional harness (mirroring OutputWAV::addWriteJob) proves queue-full rejection under tryAddJob() still leaves the buffer retryable with no leak/false pending state, plus a "buggy" counterpart proving the check discriminates a regression of the original defect class. All four host self-checks (adts_timing_self_check, run-speed-modifier-self-check.sh, wav_header_selfcheck, JobQueueContractSelfCheck) pass with -Wall -Wextra -Werror -fsanitize=address,undefined. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on #19 (
4ad5108); merge #19 first. The recording changes are directly based on that head, so the diff will collapse to the focused recording change when #19 lands.The recording path added in
d74ec00/d2018b6treated an attached splitter output as success, spun when recorder buffers filled, ignored short SD writes, and closed before queued finalization was proven. This keeps the post-mixOutputSplittertap but adds explicitIDLE/STARTING/RECORDING/STOPPING/COMPLETE/FAILEDstatus, concrete SD/open/write/finalize/overrun/queue errors, and written-byte, duration, dropped-byte, retry, and file-validity reporting.Live start only dispatches a request. The audio task opens the temp file, then initial WAV seek/header setup advances asynchronously through scheduler queue/wait states while mixer/I2S output continues; the recorder is attached only after the zero-length header is confirmed. Existing decoders/encoders retain the scheduler's legacy blocking
addJob()contract;OutputWAValone uses zero-timeouttryAddJob(), so recording queue pressure is nonblocking without changing playback job semantics. Initial and final header queue rejection stays pending and retries up to 32 times; exhaustion reports an error withfileValid=false. Buffer exhaustion or short/full/removed-SD writes detach recording without stalling playback. Immediate/repeated start-stop, rejected stop dispatch, queued teardown, failed restart validity, and duplicate attachment are guarded explicitly.Validation:
arduino-cli compile --fqbn cm:esp32:jaydDeliberately unchanged: firmware naming/UI, AAC conversion, WiFi,
DjSession, automix, the existing per-write job allocation model, and synchronousSD.exists/remove/openprimitives. Those calls contain no explicit wait loop here; moving filesystem open itself off the audio task requires a broader serialized FS scheduler API.