Merge pull request 'Issues 296, 299, 300: resolved, each with its replay' (#183) from issues/296-299-300-resolved into main

2026-10-07 21:55:20 +00:00
parent ded2173308 787f428e7e
commit 69ab2fd7f7
3 changed files with 38 additions and 6 deletions
@@ -1,8 +1,9 @@
 ---
-status: located
+status: resolved
 opened: 2026-10-07
 located-in: [mesh-controller cmd/mesh-controller (release_plan.go, watchdogs.go, status.go)]
-fixed-by: mesh-controller PR #119
+fixed-by: novox/mesh-controller PR #119
+replay: R296
 amended-design:
 ---

@@ -41,3 +42,12 @@ and the plan's revision does not move until something about it does.

 At resolution this core issue names its replay, or says in `replay-none:` why none is possible
 (ADR 0237).
+
+## Resolved — 2026-10-07
+
+Fixed by novox/mesh-controller PR #119, merged and live. The replay is R296 in mesh-lab's replays
+register (novox/mesh-lab PR #58), a test in the controller (novox/mesh-controller PR #124) laid over the
+merge before the fix. A plan in its tier since 19:48:41, read at 20:30:00 and saved four seconds before,
+reads 41 minutes and LATE; and a plan whose only build waits is not saved again by three advances. On the
+commit before the fix the line read "building for 4s" and the plan's revision moved. mesh-lab's prover
+says it is proved: it fails on the merge's first parent and passes on the fix's merge.
@@ -1,8 +1,9 @@
 ---
-status: located
+status: resolved
 opened: 2026-10-07
 located-in: [mesh-controller cmd/mesh-controller (probes.go, the self-check's D3)]
-fixed-by: mesh-controller PR #121
+fixed-by: novox/mesh-controller PR #121
+replay: R299
 amended-design:
 ---

@@ -50,3 +51,14 @@ A false urgent condition is not harmless. It trains the reader to ignore the boa

 The trail is in [01-diagnosis.md](01-diagnosis.md). This is a core issue: at resolution it names its
 replay, or says in `replay-none:` why none is possible (ADR 0237).
+
+## Resolved — 2026-10-07
+
+Fixed by novox/mesh-controller PR #121, merged and live. The replay is R299 in mesh-lab's replays
+register (novox/mesh-lab PR #58), a test in the controller (novox/mesh-controller PR #124) laid over the
+merge before the fix. Through D3 itself, with the seats read back from the store and a holder on record
+on every machine heard from, answering nothing: no holder of a seat whose verbs are all optional is said
+silent. It replays `node-uplink` while its verbs are optional, and beside it a seat whose verbs are
+optional by definition, so the outcome is still held once the seat's verbs are required. On the commit
+before the fix both were said silent. mesh-lab's prover says it is proved: it fails on the merge's first
+parent and passes on the fix's merge.
@@ -1,8 +1,9 @@
 ---
-status: located
+status: resolved
 opened: 2026-10-07
 located-in: [mesh-controller cmd/mesh-controller (upgrades.go, the merge handler)]
-fixed-by: mesh-controller PR #122
+fixed-by: novox/mesh-controller PR #122
+replay: R300
 amended-design:
 ---

@@ -48,3 +49,12 @@ new module unregistered turns each new module into two acts by hand, and the pla

 The trail is in [01-diagnosis.md](01-diagnosis.md). This is a core issue: at resolution it names its
 replay, or says in `replay-none:` why none is possible (ADR 0237).
+
+## Resolved — 2026-10-07
+
+Fixed by novox/mesh-controller PR #122, merged and live. The replay is R300 in mesh-lab's replays
+register (novox/mesh-lab PR #58), a test in the controller (novox/mesh-controller PR #124) laid over the
+merge before the fix. A catalogue merge whose changed files are only a new module's directory asks the
+build seat for that directory, at the branch merged into, and opens no plan. On the commit before the fix
+it asked for nothing. mesh-lab's prover says it is proved: it fails on the merge's first parent and
+passes on the fix's merge.