Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions AGENTS.md
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
> **LATEST CROSS-THREAD STATUS (2026-10-31, main after #234):** Read [M1/M2 & release reconciliation](docs/RECONCILED_PROGRESS_M1_M2_20261031.md) FIRST. DEVICE2+SOURCE2 are merged together in #234 (`7db4ab70`); #232 closed superseded, never merge its old head. #235 M2 FX/BUS is OPEN with Chromium regressions and not production; #229/#225 FactoryFX overlap and need selective reconcile. The native Windows FAT ZIP from `main@a7fb6151` predates #234; use a newer successful main artifact when published. No MAONO/WASAPI hardware FAT or #15 zero-VST crackle clearance, and FactoryFX approved lock still empty. Older AGENTS milestones below are historical if they conflict. Preserve one JUCE device owner, one native graph/route authority, stable L2/RAIL3, and explicit native ACK / output safety.

> **DEVICE1 CUSTOM DEVICE SETUP MERGED #228 (2026-10-31) — CI PASS, PHYSICAL FAT PENDING:** Read [PRD](docs/PRD_CUSTOM_AUDIO_DEVICE_SETUP_20261031.md), [ARD](docs/ARD_CUSTOM_AUDIO_DEVICE_SETUP_20261031.md) and [per-file handoff](docs/HANDOFF_CUSTOM_AUDIO_DEVICE_SETUP_20261031.md) before touching AUDIO DEVICE / DEVICE-DRIVER / ASIO / WASAPI. Existing WebView2 Audio I/O Matrix must become the SINGLE real device setup UI: backend, actual capture/playback endpoints, sample rate, buffer, enabled pins, APPLY with verified JUCE readback, fail-closed muted rollback. Keep ONE existing AudioDeviceManager, native routing epoch/ACK, MAIN/MONITOR/OUTPUT1/FX and #15 gate. Now DEVICE/DRIVER in the existing WebView2 matrix selects actual JUCE-reported ASIO/WASAPI/DirectSound backend, input/output endpoints, rate, buffer and verified pin masks; its native APPLY, muted rollback, readback and requestId ACK were merged in #228 at e0eaeafe88b9fdc997f3f6e7856331679db1ba7b. The old AudioDeviceSelectorComponent remains only a C++ engineering fallback and is NOT the normal UI. Physical MAONO/WASAPI FAT, vendor driver hangs and #15 0FX crackle remain UNQUALIFIED. Driver scans/opens may block vendor code: don't claim async bridge eliminates freeze; audit pinned JUCE thread affinity and instrument physical hangs. No second hardware setup UI, manager, audio graph or naive per-dropdown device restart.

> **SOURCE1 + SETUP1 — 2026-10-31:** See [source/real-device integrated milestone](docs/SOURCE1_NATIVE_MUTED_FAST_COMMIT_20261031.md). 64-pin real JUCE ASIO/WASAPI/DirectSound selector; current hardware Input→CH from MIX ROUTE auto-submits via ONE native routingDryApply/ACK only if outputs already MUTED and no other drafts. Armed output fails closed, no auto-mute/unmute; true RT clickless source crossfade and persistent named aliases are NOT IMPLEMENTED. Preserve #223 AUX live, MAIN/MONITOR/OUTPUT1, L2/RAIL3, VST/Show and physical #15 gate.\n\n> **M1 LIVE AUX POST FADER (2026-10-31):** [Native implementation & FAT contract](docs/M1_LIVE_AUX_SEND_FADER_20261031.md). Ordinary gain on an EXISTING native CH→AUX POST edge is a realtime atomic target with callback-only <=64-sample smoothing. Control write via `routingAuxSendGain` is fenced by actual device epoch, published routing revision, existing edge, 0..1 value and native ACK. It does NOT MUTE outputs, recompile topology, detach, instantiate plugins or change MAIN fader. New send/return enable, PRE/PAN and physical I/O remain structural DRAFT→MUTE→APPLY→native ACK. No fake dynamic CH11, Master FX or hardware crackle certification. Preserve source and standalone mixer-io JS parity, L2/RAIL3, DYN1B, OUTPUT1, factory lock and P0 hardware #15.
Expand Down
45 changes: 45 additions & 0 deletions docs/RECONCILED_PROGRESS_M1_M2_20261031.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
# ArSonMiX — M1/M2 release & PR reconciliation (31 October 2026)

**Canonical decision record, snapshot only.** Baseline reviewed: `main@7db4ab70a4c971487e0b921df5a4e8cc805f0183`, after #234 merge. Previous cross-thread product truth documents reflect earlier milestones and **must not overrule this newer factual ledger**. Refresh real CI and main HEAD before decisions. ArSonMiX is a low-latency *realtime Windows digital mixer* for physical or virtual I/O, live karaoke, stage, podcast and livestream — **not** a timeline editor, WebAudio engine, OBS-only mixer, or zero-millisecond hardware promise.

## 1. Source of truth: MERGED and software-verified

| Area | Established result / authority | What this does NOT prove |
| --- | --- | --- |
| Physical interface/device setup | #224 (selectable 64 JUCE pins), #228 (actual custom WebView2 ASIO/WASAPI/DirectSound driver/endpoints/rate/buffer/pin setup, muted native APPLY, ACK, safe rollback), #234 (DEVICE2 cached inventory + explicitly guarded REFRESH + exact active hardware mask/rate/buffer readback) | No proven guarantee against vendor-driver in-process hang; 64 selectable hardware pins is not 64 live mixer strips |
| Channel source | #218 native input matrix; #224 muted fast SOURCE; #233 selected-channel SOURCE2 directly chooses verified capture pins with one native route ACK; #234 contains DEVICE2 **and** preserves #233 | No universally clickless live-ARMED source switching; no arbitrary simultaneous multi-interface capture; no stable named logical source registry |
| Live mixer and bus | Current one JUCE callback/accepted routing graph: **10 input strips + 4 internal AUX returns**, native sends and AUX master, MAIN, independently selectable MONITOR hardware pair and optional secondary output with clock bridge (#203/#209/#223/#231) | No user-created CH11+, full arbitrary BUS/MATRIX/DCA/CUE, dynamic graph cutover or automated latency/PDC proof |
| GUI | Existing L2/RAIL3 compact 11-strip console and one WebView2 AUDIO I/O matrix; selected CHANNEL focus and SOURCE/AUX workflow, no second audio setup window | Not yet comprehensively qualified on low-end iGPU, vendor hardware or large plugin loads |
| VST3 and Shows | Existing native host with CH1 legacy FX and routed CH/AUX insert/Show leasing where accepted | No common CH1..N VST ownership/MASTER insert equivalence, global Studio One feature parity or zero plugin-added latency |
| Build & source audit | #234 exact-head Chromium [41173141028](https://github.com/masarray/arsonmix/actions/runs/41173141028) **SUCCESS**, Windows native/CTest/verified ZIP [41173140873](https://github.com/masarray/arsonmix/actions/runs/41173140873) **SUCCESS**; merged commit above | CI is NOT a physically audible MAONO or WASAPI FAT |
| Downloadable FAT before #234 | Last verified available `main@a7fb61511ffe27fdef1f9923695daf6899646dff` Windows x64 artifact at [41166662476](https://github.com/masarray/arsonmix/actions/runs/41166662476), artifact `ArSonMix-GUI10CH-P012-Verified-Hardware-FAT-Windows-x64` | This ZIP **does not yet contain #234**; use a successful **newer main** Windows artifact when available |

**Native invariants:** ONE `AudioDeviceManager`, ONE native callback, ONE routing authority (`web/monitor.js` + native epoch/revision/ACK), SINGLE WebView2 matrix, no JS DSP, explicit structural MUTE → APPLY → native readback/ACK, never auto-arm unknown output. Fader, mute/solo, pan and already-built AUX send gains are ordinary RT control operations, not forced topology rebuilds. Do not silently substitute a different microphone, speaker endpoint, virtual pin or clock.

## 2. Canonical PR ownership — do not double-merge

| PR | Current decision at this snapshot | Next gate |
| --- | --- | --- |
| [#234](https://github.com/masarray/arsonmix/pull/234) | **MERGED**. Integrates DEVICE2 #232 onto selected-channel SOURCE2 #233 on main. No new DSP manager; future device edits start here | Re-verify main CI + real hardware |
| [#232](https://github.com/masarray/arsonmix/pull/232) | **CLOSED, NOT MERGED** after #234. Historical old-head source; superseded and should never be bulk-merged | None; only retain tests/evidence |
| [#235](https://github.com/masarray/arsonmix/pull/235) | **OPEN, NOT YET ACCEPTED**. M2 proposal unifies FX rack presentation across CH1–10, adds bounded subgroup/FX BUS *using spare existing native CH*, still separates CH1 legacy and routed VST ownership | Fix exact-head Chromium smoke; earlier runs failed on old static assumptions and on `tests/ui_smoke.py` FX-picker visibility/count (latest investigated failure line 1464 in run 41173956993). Reconcile/rebase over #234, then exact-head UI + Windows CTest + hardware FAT. **No premature merge** |
| [#229](https://github.com/masarray/arsonmix/pull/229) | **OPEN, STALE BASE**. Reconciles FactoryFX supplier/Comp packaging, docs and PR backlog with more recent upstream than #225; not related to DEVICE or ROUTE graph | Rebase only nonoverlapping FactoryFX admission/watcher/test/docs from current main. Require approved legal SHA lock and offline test; no auto-shipped FX |
| [#225](https://github.com/masarray/arsonmix/pull/225) | **OPEN, STALE**. Original ArSonComp supplier watcher/staging integration overlaps #229 | Never merge both old branches as-is. Prefer a single reconciled FactoryFX owner on current main; close the redundant PR only when replacement is actually merged |
| Old PR #128/#117/#115/#113 and pre-M1 branches | Historical prototype work, not authority over current merged native one-graph/session architecture | Quarantine; selectively audit missing requirements, never bulk-merge divergent `Main.cpp`, `monitor.js`, FX/GUI |

## 3. Defects / high-value next milestones

**P0 — Physical signal FAT and issue #15.** Zero-VST crackling remains unresolved. On real Windows hardware test **0FX first** at driver-reported 44.1/48 kHz and 64/128/256 if available; inspect input vs output discontinuity, callback gap, true audible dropout, latency and loopback; preserve source-labeled audio report + MARK CRACKLE. Test at conservative headphone volume. Real ASIO registration, pin enable masks, default rate, MAIN/MONITOR/OUTPUT1, disconnection and accepted native route ACK all need operator evidence. An artifact ZIP is software-tested, not sound-tested.

**P1 — Finish M2 correctly, not cosmetically.** First make the CH1–10 shared FX presenter and real BUS/FX BUS staged wizard pass Chromium without breaking normal CH1 VST ADD, EDIT, BYPASS, drag order, VST lease or Show. A proposed BUS consumes a genuinely unused one of ten native channels, never fabricates CH11; next major architecture step is unified dynamic `ChannelId` realtime graph and VST ownership/PDC, separately gated.

**P1 — Driver responsive architecture.** #234 avoids scanning all vendor drivers on every Audio Setup open, but explicit refresh or actual driver switch remains synchronous JUCE vendor code; a worker on the same manager is not inherently safe. Only after measuring field stalls evaluate isolated **inventory** helper process or separated host architecture. Avoid a naive async wrapper or claim "zero freezing."

**P2 — Factory VST3 distribution.** `config/factoryfx-approved-releases.json` currently contains **`plugins: []`**. Until auditable source release, exact SHA256, publisher permission, VST class IDs, host qualification and compatible packaging are supplied, do not present ArSonComp/EQ/Verb as shipped by default.

## 4. Acceptance and CI economy

1. One functional owner per domain (DEVICE: #234; SOURCE: merged #233; M2 console/rack/bus: #235; FactoryFX: reconcile #229/#225), one accepted main state; rebases onto current main, never resurrect stale branch engines.
2. Source/quick tests first; use a **single stable head** for full Windows MSVC/native CTest/package + Chromium run. Do not repeatedly amend a PR while Windows compile is underway or start duplicate workflows.
3. Merge only after exact-head green UI/native gates, no conflicts with newly merged source/device owners, and clear negative/fail-closed tests. Then main publishes the verified user-downloadable Windows artifact.
4. Physical release remains gated by ASIO/WASAPI FAT and #15 even after all green software CI.