Give ten uncited decisions their homes in the designs they change (issue 504) #324

Merged
mesh-admin merged 1 commits from fix/504-homes-for-uncited-decisions into main 2026-10-12 11:45:29 +00:00
Contributor

The knowledge base's cited rule fails for ten accepted decisions that nothing cites: ADR 0303, 0306–0312, 0314 and 0315 (condition module.mesh-knowledge.<control-node>.unhealthy). Each design that owes the decision now names it in decisions: and states it in prose; each record then gets a progressive insight naming its design under homes.

ADR Home
0303 several modules may declare one package to-be 42 (after Who writes what)
0306 walks beside each other judge only their own builds to-be 47 (The delivery budget)
0307 the runtime serves a key's kept values to-be 32 (A module declares state too)
0308 a walk's messages are held and told once to-be 46 §8, new While a delivery walks
0309 a shared mesh-wide setting before the default to-be 27 (after the ADR 0262 paragraph)
0310 the node-reads seat to-be 41, new WP8
0311 pinned third-party images copied and checked to-be 51, new Pinned third-party images
0312 a read-write access writes through the path's group to-be 27 (An operator's shared data)
0314 a condition no healer can fix opens an issue to-be 45 §7; its line under not decided removed
0315 the record is the bus to-be 06 (the record paragraph and Open)

merge-check.sh passes locally (records, index, cycle, words). Issue 504.

The knowledge base's `cited` rule fails for ten accepted decisions that nothing cites: ADR 0303, 0306–0312, 0314 and 0315 (condition `module.mesh-knowledge.<control-node>.unhealthy`). Each design that owes the decision now names it in `decisions:` and states it in prose; each record then gets a progressive insight naming its design under `homes`. | ADR | Home | |---|---| | 0303 several modules may declare one package | to-be 42 (after *Who writes what*) | | 0306 walks beside each other judge only their own builds | to-be 47 (*The delivery budget*) | | 0307 the runtime serves a key's kept values | to-be 32 (*A module declares state too*) | | 0308 a walk's messages are held and told once | to-be 46 §8, new *While a delivery walks* | | 0309 a shared mesh-wide setting before the default | to-be 27 (after the ADR 0262 paragraph) | | 0310 the `node-reads` seat | to-be 41, new WP8 | | 0311 pinned third-party images copied and checked | to-be 51, new *Pinned third-party images* | | 0312 a read-write access writes through the path's group | to-be 27 (*An operator's shared data*) | | 0314 a condition no healer can fix opens an issue | to-be 45 §7; its line under *not decided* removed | | 0315 the record is the bus | to-be 06 (the `record` paragraph and *Open*) | `merge-check.sh` passes locally (records, index, cycle, words). Issue 504.
mesh-admin added 1 commit 2026-10-12 11:25:00 +00:00
Give ten uncited decisions their homes in the designs they change (issue 504)
mesh/merge-gate pass: the change touches no module of the mesh's graph
mesh/repo-check pass: its merge-check.sh passed
mesh/delivery delivered
23ef14d0ba
The knowledge base's cited rule failed for ADR 0303, 0306-0312, 0314 and 0315: accepted,
and nothing cited them. Each design that owes the decision now names it in its decisions
and says it in prose, so the records can name these designs as their homes.
Author
Contributor

Delivery novox/hq@23ef14d0ba6b — delivered since 2026-10-12T11:45:59Z

Delivery plan — builds nothing: the change touches no module of the mesh's graph

Transitions

  • 2026-10-12 11:25 (new) → proposed (announced): novox/hq#324's head announced
  • 2026-10-12 11:25 proposed → checked (checked): the verdict names this commit
  • 2026-10-12 11:25 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
  • 2026-10-12 11:45 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move
  • 2026-10-12 11:45 published → delivered (done): nothing for a walk to move

The commit's note under refs/notes/mesh-plan keeps every transition: git log --notes=mesh-plan.

<!-- mesh-delivery:view --> **Delivery** `novox/hq@23ef14d0ba6b` — **delivered** since 2026-10-12T11:45:59Z **Delivery plan** — builds nothing: the change touches no module of the mesh's graph **Transitions** - 2026-10-12 11:25 (new) → proposed (announced): novox/hq#324's head announced - 2026-10-12 11:25 proposed → checked (checked): the verdict names this commit - 2026-10-12 11:25 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail - 2026-10-12 11:45 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move - 2026-10-12 11:45 published → delivered (done): nothing for a walk to move The commit's note under `refs/notes/mesh-plan` keeps every transition: `git log --notes=mesh-plan`.
Author
Contributor

Verdict: approve at 23ef14d0ba

Checked each of the nine design edits (to-be 06, 27, 32, 41, 42, 45, 46, 47, 51) against the current accepted clauses of its ADR (knowledge-base.get, bound 16000, per record) rather than against the PR's own wording:

  • Every added paragraph restates its record's in-force clauses faithfully, no more: ADR 0315 (to-be 06), 0312 and 0309 (to-be 27), 0307 (to-be 32), 0310/WP8 (to-be 41), 0303 (to-be 42), 0314 (to-be 45), 0308 (to-be 46 S8), 0306 (to-be 47), 0311 (to-be 51). I did not find a sentence that claims something its ADR does not decide.
  • Placement matches the PR body's table (checked by grepping each file's headings): to-be 42 after 'Who writes what', to-be 45 under S7, to-be 46 under S8 ('While a delivery walks'), to-be 27's two paragraphs where claimed. To-be 45 also drops its now-stale 'not decided' bullet for issue 409, which agrees with ADR 0314's own progressive insight.
  • Frontmatter: only updated (2026-10-12) and decisions changed on all nine files; status/layer/code untouched, so to-be stays to-be and nothing here asserts an as-is fact. No node names, addresses, paths or usernames found in the new prose.
  • merge-check.sh passed in a fresh worktree at this head (records: all checks passed; index: 6 topics, no stored list; cycle: 436 documents, chain holds; words: 28 retired words checked, none in running prose). Gitea's checks at this head are both green: mesh/merge-gate ("touches no module of the mesh's graph") and mesh/repo-check ("its merge-check.sh passed"); 0 reviews before this one.

One thing worth flagging, not a reason to hold this PR: ADR 0309's own homes insight (naming to-be 27) is not accepted — knowledge-base history shows version 7 reviewed with three findings asking for a decision, not a fact correction (clause 3 vs ADR 0164 S4/S9, and clause 4/6's module-less write vs ADR 0277 S1/S2's module-named warrant). That is a real gap in ADR 0309 itself, surfaced by the attempt to add its home. The to-be 27 paragraph this PR adds for ADR 0309 only repeats clauses 1-7 of the still-accepted version 5 (all 'in force') and does not resolve or paper over that gap, so it does not need to wait on it — but the gap itself does: ADR 0309 clauses 4 and 6 should not be treated as settled mechanism until a follow-up decision (or a changes-link to ADR 0164 S4/S9 and ADR 0277) resolves it. Worth its own issue if one doesn't exist yet.

For contrast, ADR 0311's homes insight looked stuck the same way on a first read (pending: version 14 ...), but its history shows that was only the operator's-ask channel failing twice (telegram availability, then a 3-open-asks limit), not a content objection; it was accepted at version 15. ADR 0303, 0306, 0307, 0308, 0310, 0312, 0314, 0315 all carry accepted homes insights matching this PR's table.

Verdict: approve at 23ef14d0ba6b3d6a49134db1621cade36999f5e4 Checked each of the nine design edits (to-be 06, 27, 32, 41, 42, 45, 46, 47, 51) against the current accepted clauses of its ADR (knowledge-base.get, bound 16000, per record) rather than against the PR's own wording: - Every added paragraph restates its record's in-force clauses faithfully, no more: ADR 0315 (to-be 06), 0312 and 0309 (to-be 27), 0307 (to-be 32), 0310/WP8 (to-be 41), 0303 (to-be 42), 0314 (to-be 45), 0308 (to-be 46 S8), 0306 (to-be 47), 0311 (to-be 51). I did not find a sentence that claims something its ADR does not decide. - Placement matches the PR body's table (checked by grepping each file's headings): to-be 42 after 'Who writes what', to-be 45 under S7, to-be 46 under S8 ('While a delivery walks'), to-be 27's two paragraphs where claimed. To-be 45 also drops its now-stale 'not decided' bullet for issue 409, which agrees with ADR 0314's own progressive insight. - Frontmatter: only `updated` (2026-10-12) and `decisions` changed on all nine files; `status`/`layer`/`code` untouched, so to-be stays to-be and nothing here asserts an as-is fact. No node names, addresses, paths or usernames found in the new prose. - `merge-check.sh` passed in a fresh worktree at this head (records: all checks passed; index: 6 topics, no stored list; cycle: 436 documents, chain holds; words: 28 retired words checked, none in running prose). Gitea's checks at this head are both green: mesh/merge-gate ("touches no module of the mesh's graph") and mesh/repo-check ("its merge-check.sh passed"); 0 reviews before this one. One thing worth flagging, not a reason to hold this PR: ADR 0309's own `homes` insight (naming to-be 27) is not accepted — knowledge-base history shows version 7 reviewed with three findings asking for a decision, not a fact correction (clause 3 vs ADR 0164 S4/S9, and clause 4/6's module-less write vs ADR 0277 S1/S2's module-named warrant). That is a real gap in ADR 0309 itself, surfaced by the attempt to add its home. The to-be 27 paragraph this PR adds for ADR 0309 only repeats clauses 1-7 of the still-accepted version 5 (all 'in force') and does not resolve or paper over that gap, so it does not need to wait on it — but the gap itself does: ADR 0309 clauses 4 and 6 should not be treated as settled mechanism until a follow-up decision (or a changes-link to ADR 0164 S4/S9 and ADR 0277) resolves it. Worth its own issue if one doesn't exist yet. For contrast, ADR 0311's homes insight looked stuck the same way on a first read (`pending: version 14 ...`), but its history shows that was only the operator's-ask channel failing twice (telegram availability, then a 3-open-asks limit), not a content objection; it was accepted at version 15. ADR 0303, 0306, 0307, 0308, 0310, 0312, 0314, 0315 all carry accepted `homes` insights matching this PR's table.
mesh-admin merged commit d0e23b1dbf into main 2026-10-12 11:45:29 +00:00
mesh-admin deleted branch fix/504-homes-for-uncited-decisions 2026-10-12 11:45:29 +00:00
Sign in to join this conversation.
No Reviewers
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: novox/hq#324