Repository navigation
P1J: preserve live VST3 instances through routing edits + native session SSOT + Signal Flow routing - #130
Conversation
…o-lock RT processing and deferred retirement
…cross graph publishes
…tural APPLY; report reuse in ACK
…nd expose editor through Signal Flow
…showing conceptual preview
…e routing panel entry
…e graph swap; retirement and RT remain safe
…Flow editor entry
…, precise FAT and remaining gates
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. |
…dor releaseResources
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ffaec7db7f
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".
| processing.clear(std::memory_order_release); | ||
| } | ||
| }; | ||
| using FxBank=std::array<std::shared_ptr<FxInstance>,channels>; |
There was a problem hiding this comment.
Update the existing contracts for shared FX ownership
The shared-instance refactor leaves the existing P1H contract incompatible with the implementation: node --test tests/live_routing_ui_contract_test.mjs fails at line 91 because it still requires std::array<std::unique_ptr<Processor>,channels>, and the following assertion also requires the removed direct processors_[strip]->processBlock(buffer,midi) call. Adding the P1J test does not replace those assertions. The Validate native UI JavaScript step in .github/workflows/p1-windows.yml runs this suite and exits on failure before configuring the native build, blocking the required Windows validation. Update the older assertions to verify independent strip instances through the new shared-instance wrapper.
AGENTS.md reference: AGENTS.md:L26-L26
Useful? React with 👍 / 👎.
…uring graph retirement
…mutate juce::var in deviceStopped callback
P1J — stable vendor FX across routing APPLY + native-acknowledged routing SSOT + Signal Flow access
Based on current main@abcdf2b9 (P1I 10-channel native faders/meter/PAN/MUTE/SOLO and latest merged P0 GUI/Audio Setup/Snapshot work). Explicitly preserves user-approved CH/MASTER GUI, 30fps LED scheduler, Orbit, normal CH1 standalone FX, hardware port identity and zero-FX diagnostics.
Actual native milestone
FxInstanceowns its MIDI scratch; immutable routing snapshots share it via control-thread-onlyshared_ptrcopies. No audio callback allocation, locks, refcounting or spin. Atomic try-enter gates unexpected concurrent vendor access, fails closed. Vendor C++ exceptions silenced; OS/SEH crash isolation not claimed.RealtimeRoutePublisher::controlPrepared(epoch,revision)exposes the active immutable FX identity only to single control writer; old snapshot leases retain lifetime.acceptedRoutesaved only after publish ACK; status returns current physical pins, source/BUS assignments, FX IDs and internal PRE/POST sends, plus actual live channel controls. On opening the editor, JS hydrates from the authoritative accepted graph (not stale browser draft), then overlays current native fader/PAN/MUTE/SOLO.fxReusedvisible in ACK.window.__arMixOpenRoutingLab, retaining old standalone preview; no toolbar/mixer redesign.docs/P1J_STABLE_FX_ROUTING_SESSION_20261029.md.No false equivalence: acceptedRoute is runtime-only (not portable Shows), 1 FX/channel, BUS consumes strips, 2 hardware outputs, PDC absent, output must be muted for structural APPLY, #15 hardware crackle FAT still open. One-SHA exact-head Windows/MSVC CTest/packager and Chromium smoke MUST pass before merging. Do not touch stale parallel #115/#117.