Repository navigation
M3-B reconciliation: extract ChannelId readback projection from ENGINE #266 into approved integration seam - #402
Conversation
…explicit callback admission and CTest
|
M3-B cross-thread reconciliation: I am extending this existing INTEGRATOR PR only, not opening a competing PR. I will add an isolated, fail-closed v8 |
…elId PCM operations (fail closed)
M3-B native control ABI handoff — REAL code, single INTEGRATOR PR #402Commit 4163bf3faebf was added to existing PR #402, without starting a second PR/workflow or changing ENGINE source.
ENGINE #266 required nextEnsure original CH1–10 + dynamic INPUT ChannelIds actually mutate the one installed PCM callback via This keeps M3 progressing within one owner branch and avoids claiming First Real Sound from C++ fixtures alone. |
…; keep unadmitted operations fail closed
M3-C actual INTEGRATOR runtime stitch — commit 427d62466dcbOn existing #402 only. No duplicate GUI/engine/CI workflow. The previous exact-head Windows run 41910167994 compiled the Windows app and 55 CTests; 54 PASS, 1 FAIL at our own newly-added assertion New integrated changes:
ENGINE #266 remains bottleneck: current |
M3-C INTEGRATOR checkpoint — 2026-11-05, current HEAD
|
M3C ↔ M4 concrete reconciliation — M4 merged, M3C owner handoff (2026-11-05 WIB)INTEGRATOR-owned M4 #404 was merged as #402 HEAD Please reconcile #402 against exact current main using its existing branch, preserve only its unique code/tests ( Combined ABI verified from source: M3C projector emits canonical decimal-string ChannelId/EdgeId, Acceptance for joined PR #402: exact-current-main merged tree, Windows full CTest, old MAIN removed on Route Only, exact accepted tap/gain, no double-dry, muted/solo/fader input→MAIN proof from installed callback, original VST3/Show/AUX retained, GUI exactly one /preview/, hardware crackle #15 and FAT still stop-ship. Merge only after joined proof; no second engine/GUI/release workflow. Please post current #402 HEAD + joined CI evidence back on #397 before merging so independent threads can use one authoritative SHA. |
…ary-route/upstream double-dry admission
M3-D joined INTEGRATOR reconciliation — actual commit cb7a44ebf104Control tower #397 · sole INTEGRATOR PR #402 · baseline main Two still-open #404 review bugs addressed in code + 3 regression cases:
Current exact-head joined CI: Windows 41921060410 + Chromium 41921060409 on ENGINE #266 remains sole PCM callback owner (HEAD Current application is NOT FIRST SOUND: |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
M3-D exact-head outcome / deliberate fail-closed PR stateCurrent INTEGRATOR #402 HEAD Implemented and verified code: previous joined Latest a2a admission: Chromium #41922794387 SUCCESS, including mandatory IMPORTANT: not First Real Sound. G0 |
|
FAT #406 encountered unrelated INTEGRATOR-owned M2.6 Windows CRLF regression: tests/workstream_coordination_contract.test.mjs line225 splits workflow content on LF only and fails |
INTEGRATOR handoff: P0 CI unblock now landed on mainAccepted M3 #402 remains the sole native INTEGRATOR owner (last observed head No sound qualification claim: real PCM callback #266 + actual GUI native ACK/snapshot + physical FAT #15 remain product milestones. Ruleset administrator still needs to require new |
… with protected Product Qualification main
M3-E INTEGRATOR active execution — protected main + exact owner handoffs (2026-11-05 WIB)WORKSTREAM: INTEGRATOR. Baseline ENGINE #266 owner handoff REQUIRED for actual FIRST SOUND: Branch still FAT #406 owner handoff: reconcile FAT-only ZIP/collector/evidence kit from Owner / merger must never replace frozen approved https://masarray.github.io/arsonmix/preview/ skin. End goal: one accepted EXE commit, one native callback, ACK→later accepted PCM snapshot, measured evidence. |
|
@codex review M3-E exact-head review request for sole INTEGRATOR PR #402 |
|
Codex Review: Didn't find any major issues. Already looking forward to the next diff. Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c99f60233a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| return reject("INVALID_PAN_VALUE"); | ||
| acceptedPan=requested; | ||
| } | ||
| const double gain=acceptedDb<= -60.?0.:std::pow(10.,acceptedDb/20.); |
There was a problem hiding this comment.
Preserve native gain when translating a pan-only intent
When an admitted channel has native gain 0.001, the projection reports faderDb: -60 and the readback consistency check passes. However, translating SET_PAN then converts that unchanged fader value to gain 0, so moving the pan also silences the channel. The same round-trip can alter gains at the upper projection clamp. For SET_PAN, retain the already validated nativeGain; apply the dB-to-gain conversion only for SET_FADER.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
Addressed in 039b87462153 on existing INTEGRATOR #402. SET_PAN now forwards the validated native linear gain unchanged; only explicit SET_FADER recomputes it from dB. Native projection/translator regression covers gain 0, 0.001 (-60 dB nonzero) and 4.0 (+12 dB clamped), checks PAN=-0.35 preserves gain, and verifies explicit fader -60 still mutes. No ENGINE/GUI/FAT paths touched. Awaiting new exact-head Windows CTest / Product Qualification before resolving this thread/merging.
|
@codex review Please review the latest INTEGRATOR head |
|
Codex Review: Didn't find any major issues. Can't wait for the next one! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Meaningful milestone — integration of real ChannelId producer readback without another engine
Base: current main at branch creation
8ade8df7e151c421c736ff1c87ed9b0f2e7eef61(#396 + #398 + #400).Reused implementation:
src/ApprovedNativeV8Projection.hfrom sole ENGINE #266. Selectively extracted asintegration/ApprovedNativeV8Projection.hunder INTEGRATOR ownership; retains canonical uint64 ChannelId/EdgeId mapping, native fader dB/peak/insert/sends projection into approved public v8 schema.Hardening: parser requires explicit installed JUCE ChannelId callback admission. A plausible bus-shaped JSON blob is not proof of production PCM and G0 never promotes it; malformed or stale source leaves snapshot unchanged.
productionCallbackLiveremains false until real callback and FAT criteria. Noweb/console-native.*, alternate GUI, new audio callback, graph, or native DSP.Regression: adds JUCE CTest
M3BNativeV8StableIdReadbackAndFailClosedProjection: uint64 beyond JS safe integer, valid MAIN/edge/fader, native capability only with explicit admitted callback, rejected fake/bad bus, missing MAIN, stale epoch/duplicate IDs and no mutation on failure. Wires into existing Windows mixer-routing build dependency, no new workflow.Deliberately not included: no
src/Main.cppdispatch orsrc/LiveAudioEngine.hPCM changes. ENGINE #266 is sole real callback owner and must hand off genuine acceptednativeBus+ installed callback provenance to INTEGRATOR before final wire-up. #266 should use this integration path rather than merge its historicsrc/ApprovedNativeV8Projection.hand should NOT cherry-pick obsolete CMake/old UI.Acceptance: exact-head Windows JUCE EXE + CTest pass; source provenance and Chromium unaffected. Physical #15 FAT still PENDING. This PR is one integration seam, not full Milestone A or user UAT.