Repository navigation
refactor(gui G-MV28): unify verified M4 FX rack/slot transactions and stale ACK fencing - #814
Conversation
… ACK/readback policy
…CK into FX VM authority
…, driver remap and one bridge adapter
Codex Review Summary: ✅ CompletedThis 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. |
Deploying arsonmix with
|
| Latest commit: |
c862ea5
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://4a5cdbf4.arsonmix.pages.dev |
| Branch Preview URL: | https://refactor-g-mv28-m4-fx-author.arsonmix.pages.dev |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f29f4740a8
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".
|
@codex review |
|
/arsonmix-ci ui c862ea5 |
|
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". |
G-MV28 final handoff — MERGED MAIN on 2026-11-19Verified GitHub state: PR #814 merged/closed from exact reviewed head Hasil untuk pengguna: M4 FX Rack header and per-insert bypass now use a SINGLE verified native-authority decision and request adapter. A stale/retained/disconnected M4 session, legacy-vs-installed route mismatch, unsafe host epoch/revision token, missing FX lease or concurrent pending action cannot send an unrelated physical FX power command. Rack ACK and independent readback both fence exact producer epoch/revision. Slot bypass and full-rack bypass remain separate atomic operations and preserve each plugin's deliberately stored bypass flag. Actual ownership removed: obsolete Changed files (3):
Exact-SHA evidence:
Scope/safety: Approved Owner-facing completion: STATUS MERGED MAIN, CODE TESTED; AMAN DIPAKAI = YA as protected source GUI baseline; physical Windows ASIO/hardware behavior BELUM TERUJI on this commit. TINDAKAN USER = TIDAK PERLU MELAKUKAN APA-APA. NEXT G-MV29 should audit actual remaining duplicated physical WIRE/FX/Device authority, extracting only responsibilities that can be deleted from bridge, and preserve all approved GUI and PCM invariants. |
G-MV28 — one verified M4 FX Rack/slot power authority and transaction adapter
Problem sourced from live
mainafter merged G-MV27 #811:requestM4RackPowerpreviously dispatched a physicalroutingFxRackBypasssolely whenmixerSession.authority === "M4"and two safe-looking session revision counters existed, without checking verified M4 producer/readback, retained/disconnected state, installed device identity, or that the legacy M4 producer revision was actually the same. Slot bypass checkedm4Verifiedand was dispatched by a different bridge branch. Rack ACK matching was also duplicated in the bridge. A stale session or same-epoch remap could therefore produce a wrong-owner native rack command.Implemented:
fxRackVM.m4PowerPlanpure selector now owns the single M4 rack and insert power admission: exact accepted snapshot,admitted === false, not retained/offline, M4 authority, same native device epoch, same accepted session/legacy route revision, active/accepted physical M4 producer, verified 0–9 native lane, exact 0–7 FX slot and instance ID, boolean bypass state, per-control pending lease and strictly JS-safe host IPC tokens (never round uint64).requestNativeFxPowerandrequestM4RackPowerwith onerequestM4FxPoweradapter. It only packages the VM-produced action/payload into the unchanged nativeroutingFxBypass/routingFxRackBypassactions, retains separate rack/slot pending maps and existing respective 5s/8s bounded timeout, ACK and independently installed readback. No second dispatcher/controller.m4RackAckResulthandles exact ACK matching in the pure ViewModel;m4RackSettlementnow requires the full accepted native snapshot, rejecting changed authority, retained graph, missing actual M4 producer, mismatched legacy revision or epoch rather than trusting only a session fragment..smixedits, FX preset state, DSP/PCM/VST3, CSS/DOM, approved/preview/, global FX A/B, WIRE/SOF and N-MOD native PRs refactor(native N-MOD-14–18): durable Player transactions, library plans and state projection #799/refactor(native N-MOD-19): isolate DIRECT recovery coordinator and fault evidence #813.Checks required before merge: Fast UI + Fast Native on exact PR head, Codex review with all findings resolved, one exact-SHA Full Chromium
/preview/all steps, required fail-closed merge gate. No Windows installer or ASIO hardware FAT requested for this pure GUI/bridge change; physical audio qualification remains separate. Never merge pending/failing statuses or replay N-MOD work.Based on
main@3fc172da943352ba2f4952279f53fa9ea3b7f228. Pure policy smoke and JS syntax passed before PR.