pacman's mirror list keeps a fallback tail; a failed check leads the merge ask's headline (ADR 0586, issues 1876, 1166) #589

2026-11-01T17:09:19Z

mesh-tools commented

2026-11-01 17:09:19 +00:00

Two small changes the operator approved on 2026-11-01 18:03, one check, one Approve. Issues 1876 and 1166. Decision: ADR 0586 (changes ADR 0585 §2).

(1) pacman's shared mirror list keeps every mirror at the newest last sync first, then a fallback tail of up to four of the next freshest; mirrors whose last sync reflector cannot tell (it prints them as the current time) are left out; pacman_mirrors names the freshest and the tail. Why: the literal rule usually kept one mirror (last syncs minutes apart), so one mirror down failed a whole weekly round.
(2) the merge ask's headline starts 'Check failed — merge into main: …' when the check failed, before the cut to the headline length. Why: merges are not gated on the check (the operator's choice of 10-27), and on 2026-11-01 a failed-check ask read exactly like a passed one.

Quick vs thorough: both are the one place each lives (Freshest; the gate's composeMergeAsk is the only builder of the headline, the teller relays it), so quick is the whole fix. Tests: rows in existing tables, no test count rise; unit, go vet -tags integration and the offline build pass in pacman and mesh-delivery-gate. Next after delivery: ace's full upgrade, which the operator allowed once this is in.

Reviewed-by: reviewerP@shanks at 90f7f87c2a
Pushed-by: lead-47@shanks

https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL

Two small changes the operator approved on 2026-11-01 18:03, one check, one Approve. Issues 1876 and 1166. Decision: ADR 0586 (changes ADR 0585 §2). (1) pacman's shared mirror list keeps every mirror at the newest last sync first, then a fallback tail of up to four of the next freshest; mirrors whose last sync reflector cannot tell (it prints them as the current time) are left out; pacman_mirrors names the freshest and the tail. Why: the literal rule usually kept one mirror (last syncs minutes apart), so one mirror down failed a whole weekly round. (2) the merge ask's headline starts 'Check failed — merge into main: …' when the check failed, before the cut to the headline length. Why: merges are not gated on the check (the operator's choice of 10-27), and on 2026-11-01 a failed-check ask read exactly like a passed one. Quick vs thorough: both are the one place each lives (Freshest; the gate's composeMergeAsk is the only builder of the headline, the teller relays it), so quick is the whole fix. Tests: rows in existing tables, no test count rise; unit, go vet -tags integration and the offline build pass in pacman and mesh-delivery-gate. Next after delivery: ace's full upgrade, which the operator allowed once this is in. Reviewed-by: reviewerP@shanks at 90f7f87c2ac180d3087659c5d229849b76f6e42f Pushed-by: lead-47@shanks https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL

mesh-tools added 2 commits 2026-11-01 17:09:22 +00:00

Keep the freshest mirrors first and a tail of four more in pacman's shared list, so a node has a fallback (ADR 0586, issue 1876) 68b5febee7

Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL

Say a failed check first in the merge ask's headline, since a merge is not gated on the check and the headline is the phone's only warning (issue 1166)

mesh/check being carried to 90f7f87c2ac1 on ce3bedd584b8 | from 3483f603ae3d

mesh/delivery merged: source.main moved to novox/mesh-catalog@ce3bedd584b8

Details

90f7f87c2a

Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL

mesh-tools commented

2026-11-01 17:09:49 +00:00

Delivery novox/mesh-catalog@90f7f87c2ac1 — merged since 2026-11-01T18:04:07Z

Source — source.main moved to novox/mesh-catalog@ce3bedd584. Whether it reached the machines is mesh-delivery-view's, derived from source.main, the wants and the haves.

Transitions

2026-11-01 17:09 (new) → proposed (announced): novox/mesh-catalog#589's head announced
2026-11-01 17:09 proposed → ready (accepted): a verdict names this commit and the trunk's head as its trunk commit, the gate passed or warned, and the repository's own check did not fail; or the check gates no merge (check-gates-merge false)
2026-11-01 18:04 ready → merged (merged): on the trunk its modules follow: merged there

The commit's note under refs/notes/mesh-plan keeps every transition of this change up to its merge: git log --notes=mesh-plan.

**Delivery** `novox/mesh-catalog@90f7f87c2ac1` — **merged** since 2026-11-01T18:04:07Z **Source** — source.main moved to novox/mesh-catalog@ce3bedd584b8. Whether it reached the machines is `mesh-delivery-view`'s, derived from `source.main`, the wants and the haves. **Transitions** - 2026-11-01 17:09 (new) → proposed (announced): novox/mesh-catalog#589's head announced - 2026-11-01 17:09 proposed → ready (accepted): a verdict names this commit and the trunk's head as its trunk commit, the gate passed or warned, and the repository's own check did not fail; or the check gates no merge (check-gates-merge false) - 2026-11-01 18:04 ready → merged (merged): on the trunk its modules follow: merged there The commit's note under `refs/notes/mesh-plan` keeps every transition of this change up to its merge: `git log --notes=mesh-plan`.

mesh-tools merged commit ce3bedd584 into main

2026-11-01 18:04:06 +00:00

mesh-tools deleted branch fix/mirror-tail-headline

2026-11-01 18:04:06 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: novox/mesh-catalog#589