Issue 955, a third fault found by #340's delivery (playbook 08). On novox the provider logged "the provider stopped: watching mesh-controller_assignments: ... reads no state ..." at 00:44:30 and never ran again. #341 fixed the grant (delivered 01:17), but only the manifest changed: the provider's bundle digest (972d97e8) did not, so the tool runner never restarted it, and no machine's key is registered.
Service.KeepRunning (service.go): main.go runs svc.KeepRunning(ctx, p.Run); whenever Run ends it starts again, pausing 5 s doubling to 2 min, said once per distinct error, reset after a 2-min run.
Tests (service_test.go) failed first, pass now.
Reviewed (larger model), approved at 4877262 (tree 5292fb9); this forge commit carries the same diff. Opened along the way: issue 988 (each restart leaves one more bound watch handler live: a child context per start).
Issue 955, a third fault found by #340's delivery (playbook 08). On novox the provider logged "the provider stopped: watching mesh-controller_assignments: ... reads no state ..." at 00:44:30 and never ran again. #341 fixed the grant (delivered 01:17), but only the manifest changed: the provider's bundle digest (972d97e8) did not, so the tool runner never restarted it, and no machine's key is registered.
- `Service.KeepRunning` (service.go): main.go runs `svc.KeepRunning(ctx, p.Run)`; whenever Run ends it starts again, pausing 5 s doubling to 2 min, said once per distinct error, reset after a 2-min run.
- Tests (service_test.go) failed first, pass now.
Reviewed (larger model), approved at 4877262 (tree 5292fb9); this forge commit carries the same diff. Opened along the way: issue 988 (each restart leaves one more `bound` watch handler live: a child context per start).
mesh/merge-gate head b159efd65673 merged into trunk 522780d5fcfa — pass: builds gitea → novox; no bus step; every machine composes with the change as it di…
mesh/repo-check head b159efd65673 merged into trunk 522780d5fcfa — pass: its merge-check.sh passed, each part in its toolchain: merge-check.sh (go), merge-…
Deliverynovox/mesh-catalog@b159efd65673 — delivered since 2026-10-16T23:32:27Z
Delivery plan — builds gitea → novox; no bus step
tier 0: gitea
novox: receives gitea
gitea provides npm-package-registry, git, gitea-read, forge-account: its consumers are sent again after it
gitea keeps data (ADR 0233): what it holds is backed up before it moves
Machines
gitea on every machine running it: sent
Transitions
2026-10-16 23:26 (new) → proposed (announced): novox/mesh-catalog#342's head announced
2026-10-16 23:29 proposed → checked (checked): a verdict names this commit and the trunk's head as its trunk commit
2026-10-16 23:29 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
2026-10-16 23:30 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move
2026-10-16 23:30 published → delivering (go): its walk started: let go by this owner in its group's order, by a person, or on the controller's own path — not a failed walk that carried a later commit for it, after which it waits for a walk of its own commit
2026-10-16 23:32 delivering → delivered (done): its walk is done, carrying its own commit: every machine of its deploy plan passed, or was left as its policy says
The commit's note under refs/notes/mesh-plan keeps every transition: git log --notes=mesh-plan.
<!-- mesh-delivery:view -->
**Delivery** `novox/mesh-catalog@b159efd65673` — **delivered** since 2026-10-16T23:32:27Z
**Delivery plan** — builds gitea → novox; no bus step
- tier 0: gitea
- novox: receives gitea
- gitea provides npm-package-registry, git, gitea-read, forge-account: its consumers are sent again after it
- gitea keeps data (ADR 0233): what it holds is backed up before it moves
**Machines**
- gitea on every machine running it: sent
**Transitions**
- 2026-10-16 23:26 (new) → proposed (announced): novox/mesh-catalog#342's head announced
- 2026-10-16 23:29 proposed → checked (checked): a verdict names this commit and the trunk's head as its trunk commit
- 2026-10-16 23:29 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
- 2026-10-16 23:30 ready → published (merged): on the trunk its modules follow: merged there, and its walk opened — or nothing for a walk to move
- 2026-10-16 23:30 published → delivering (go): its walk started: let go by this owner in its group's order, by a person, or on the controller's own path — not a failed walk that carried a later commit for it, after which it waits for a walk of its own commit
- 2026-10-16 23:32 delivering → delivered (done): its walk is done, carrying its own commit: every machine of its deploy plan passed, or was left as its policy says
The commit's note under `refs/notes/mesh-plan` keeps every transition: `git log --notes=mesh-plan`.
Reviewed by lead-57@novox: Reviewer (opus) approved 4877262 (tree 5292fb9), the same tree as this forge commit: the provider restarts a Run that ends, with backoff; restart re-takes its watches; it touches only agent- users. Marked through mesh-delivery.mark-ready (novox/hq issue 913).
ready at b159efd65673c5c8a8c609f9b4534fb304b8e671
Reviewed by lead-57@novox: Reviewer (opus) approved 4877262 (tree 5292fb9), the same tree as this forge commit: the provider restarts a Run that ends, with backoff; restart re-takes its watches; it touches only agent-<node> users. Marked through mesh-delivery.mark-ready (novox/hq issue 913).
Merge check at b159efd6
Judged: head b159efd65673 merged into trunk 522780d5fcfa; read beside it: mesh-controller b0d489a86c23, mesh-controller-main b0d489a86c23, mesh-host cd13f7726ef0, mesh-tools 7412c5b6c9c5
mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: gitea)
mesh/repo-check: PASS — its merge-check.sh passed, each part in its toolchain: merge-check.sh (go), merge-check-typescript.sh (typescript)
Delivery plan — builds gitea → novox; no bus step
tier 0: gitea
novox: receives gitea
gitea provides npm-package-registry, git, gitea-read, forge-account: its consumers are sent again after it
gitea keeps data (ADR 0233): what it holds is backed up before it moves
Run by the build seat on shanks as build-1792193181127487471 (builds --log build-1792193181127487471).
✔ the accounts that may mark are the setting's, and only the mesh's own when it cannot be read (1.309141ms)
✔ the newest status of each context is kept, with what it said, when and by whom (20.458297ms)
✔ a commit merges only when every required status is a success; a warning blocks (0.755142ms)
✔ no rule requires nothing; a rule that cannot be read says nothing, never yes (0.306405ms)
✔ a commit is contained in another when the compare finds nothing on it the other lacks (0.947875ms)
✔ a note line is appended once, as the forge's user, in its own repository (12.451969ms)
✔ a note that cannot be written is said, never read as written (0.444877ms)
✔ one view per pull request, found by its marker; only the mesh's statuses (0.708282ms)
✔ a build machine taking the check is said on the head: building on which machine, since the local time (28.08558ms)
✔ the verdict heard while a 'building' write is in flight lands last and stays (10.853304ms)
[gitea] novox/hq@a6b08de8: the verdict of build-1791680941000000000 is an older run's than the check now pending; not written
✔ the verdict written while a start's read of the forge is in flight stays the newest status (2.976048ms)
✔ the verdict heard while a 'building' write still waits its turn: the 'building' is never written (0.672311ms)
✔ a start heard after the verdict writes nothing (0.797123ms)
✔ a start from before the check now pending (a run before a recheck, the stream's history) writes nothing (0.737832ms)
✔ an old run's verdict redelivered after a recheck's pending is not written over it (1.24762ms)
[gitea] novox/hq@a6b08de8: the verdict of build-1791680941000000000 is an older run's than the check the forge holds pending; not written
[gitea] novox/hq@a6b08de8: the verdict of build-1791680941000000000 is an older run's than the check now pending; not written
✔ an old run's verdict in flight when the recheck comes lands before the recheck's pending (4.256058ms)
✔ a rechecked head says where its new run is: building on which machine (1.056937ms)
✔ the recheck's own verdict is written, even asked a little before the pending by another machine's clock (0.805243ms)
✔ after a restart, an old run's verdict is weighed against the check the forge's history holds open (0.765172ms)
✔ a recheck the forge does not take fails its caller, and the head is as it was (0.655061ms)
✔ the status tool writes the merge check's pending through the writer, and the mesh's other statuses as they were (7.330097ms)
✔ when a check was asked is read from its id; the opening of the check now pending from the gate's history (0.272724ms)
✔ a head this process did not announce (it restarted) is said only while the forge holds its gate pending (0.368475ms)
✔ a repository as the mesh clones it is told from another (0.273354ms)
✔ gitea_api refuses the forge's administration, a branch's protection and the merge check's status, and sends nothing (5.915085ms)
✔ gitea_api refuses a file written to the default branch or one a rule guards, and writes one on a branch of its own (55.490358ms)
✔ a protection rule covers a branch as the forge decides it: classes, alternatives, a plain name in any case (issue 539) (28.805611ms)
✔ gitea_api still writes what an author writes (8.261462ms)
✔ gitea_branch_protection_set only strengthens (16.763068ms)
✔ gitea_merge_pull_request merges only a head whose required statuses all succeeded, and only that head (46.451494ms)
✔ gitea_merge_pull_request refuses a verdict judged against an earlier trunk, and asks the check once (150.821964ms)
✔ gitea_merge_pull_request merges only the head it is named, and refuses one that is not the head (17.055153ms)
✔ gitea_merge_pull_request is loud when the forge merged a head other than the one it was handed (15.068011ms)
✔ gitea_merge_pull_request says the merge happened when reading back what landed fails (26.038796ms)
✔ gitea_merge_pull_request reads back a fast-forward, and says it cannot read back a squash or a rebase (58.507426ms)
✔ gitea_delete_repo deletes only the tools' own user's repository, unguarded, untouched for 30 days, named in full (20.310085ms)
✔ gitea_api sends the path it judged, and refuses moving a pull request's base (5.279545ms)
✔ gitea_merge_pull_request merges only into a branch whose protection requires the checks (6.444373ms)
✔ a rule named for a branch is not made weaker than the glob rule that guards it now (15.122842ms)
✔ a repository owner or name is never another path (0.315465ms)
✔ the judges, without a forge (0.421917ms)
✔ main's protection is never weakened in its push or its merge whitelist (issue 948, ADR 0398 §2) (0.286835ms)
✔ gitea_api never acts as another user: sudo, in any case, and any query key it does not send are refused (3.141691ms)
✔ gitea_api never approves a pull request nor requests changes: those are the operator's (2.749674ms)
✔ no tool's caller text becomes another path or another user: `..` and `?sudo=` are refused before the forge (5.303825ms)
✔ gitea_merge_pull_request merges a verdict that read siblings only with its group passed and each sibling head reviewed (82.833885ms)
✔ a verdict put out of date by a sibling's move is refused in its own words, and asked again once (22.406428ms)
✔ every page of a pull request's files is read, though the forge caps a page at fifty (34.085295ms)
✔ a pass asks only the repositories that moved since the last look, every one before the first (0.591479ms)
✔ the directories above a merge's files that hold a module at the commit are said, the root never (14.37936ms)
✔ a lookup the forge refuses is thrown, not read as no module (5.115102ms)
✔ a renamed file is said under its old name too, as gone (novox/hq ADR 0267) (7.607402ms)
✔ a new head beyond the newest 20 open pull requests is announced, exactly once (57.942987ms)
✔ a pass pages until it reaches back past its floor, and no further (28.592758ms)
✔ gitea_branch_protection_get answers every rule, read with the administrator account the tools' token cannot read with (27.218116ms)
✔ gitea_branch_protection_get answers the rule for a branch, null when it has none (7.294806ms)
✔ gitea_branch_protection_set still gets the forge's refusal, as the tools' account (3.455335ms)
✔ the statuses asked for replace the rule's, and an empty list requires none (4.591194ms)
✔ a new rule refuses direct pushes unless asked; an administrator's override only when said (0.339095ms)
✔ main's body always refuses pushes and lets only the tools' account merge, whatever is asked (0.780973ms)
✔ another branch's body carries no merge whitelist (0.265824ms)
✔ the rule main is put right in the forge's own words: the body a setup sends carries no status checks of its own (0.333026ms)
✔ each open pull request's new head is announced once, and nothing closed or merged (24.091356ms)
✔ a verdict is the head commit's status; an error is the forge's error, never a success (1.27357ms)
✔ each layer is its own status: the gate with the modules it judged, the repository's own check beside it (0.868144ms)
✔ no verdict sets warning on a required check; only a person's decision fails (1.398602ms)
✔ a commit status is set on the commit, under the merge check's context (52.950507ms)
✔ a change plan is the gate's result: said on the status and, when it builds something, as a comment (0.648881ms)
✔ a head announced is pending in the build queue, since the local time it was announced (0.400307ms)
✔ the vault's token is used, a setting of it is ignored and said once, without its value (6.419583ms)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 at bbbbbbbbbbbb, it is now at dddddddddddd: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 was closed unmerged: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 merged onto cccccccccccc, its main is now eeeeeeeeeeee: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 was merged otherwise than its head bbbbbbbbbbbb onto cccccccccccc: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 was merged otherwise than its head bbbbbbbbbbbb onto cccccccccccc: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-host's main moved from 444444444444 to 333333333333: 0 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
✔ both statuses name the siblings the verdict read, keeping the verdict word and the start of its reason within 140 characters (2.529881ms)
✔ a verdict that could not learn its siblings says so on the status and in the comment, and names none (0.415717ms)
✔ a verdict without siblings is said exactly as before (an older controller) (0.817723ms)
✔ the siblings a verdict names are kept per member, replaced by its next verdict, and removed by one that read none (0.957835ms)
✔ a sibling's new head puts both statuses of the member back to pending, with the reason, and runs nothing (3.591048ms)
✔ a sibling closed unmerged puts the member's verdict out of date (0.433287ms)
✔ a move of the sibling's base puts it out of date, naming both commits (0.424007ms)
✔ the sibling's own merge of the head read onto the base read does not, in either order of the two merges (0.64603ms)
✔ a merge of another head, or onto another base, is a move (0.829733ms)
✔ only the members whose verdict names the sibling are touched, and a member that moved on is dropped (0.713082ms)
✔ a trunk that moved tells the siblings' weighing; one first seen does not (0.63278ms)
✔ the group rule names every missing thing: the group status at each head read, and each sibling head's review (1.168728ms)
✔ a merged or closed member's entry is dropped, so a sibling's later push never re-pends its merged head (0.266554ms)
✔ a consumer that asks nothing, or write, publishes: the package team (1.685737ms)
✔ a consumer that asks to read joins the readers, who write no package (0.534909ms)
✔ a reader is made, put on the readers team, and taken off the package team (0.674611ms)
✔ a publisher is made and put on the package team, as before (0.255234ms)
✔ first start: mints with the admin account, keeps the token at 0600, asks for the tools' scopes only (39.722356ms)
✔ second start: reuses the kept token, mints nothing (28.482886ms)
✔ the forge rejects the kept token (its data was restored): minted afresh, once, and the call goes through (30.112652ms)
✔ a token kept with write:admin by an earlier build is minted again without it at the next start, and the forge drops the old one (29.415911ms)
✔ a kept token lacking a scope the tools need: the forge refuses the call for the scope, the token is re-minted with the whole list, and the call goes through (29.945299ms)
✔ the kept file is gone but the forge still holds a token by that name: replaced, not refused (24.353419ms)
✔ concurrent first calls share one mint (13.424675ms)
✔ two tokens hold the name (two mints raced): both are deleted by their id, and one fresh token is minted (14.640755ms)
✔ two processes on one kept file, both without a token, mint once: the second waits for the first and reuses it (73.922913ms)
✔ a lock left by a process that died mid-mint is taken over, not waited on for ever (7.274276ms)
✔ the admin account is refused: said plainly, nothing kept, and the next call fails the same way rather than crashing (15.61371ms)
✔ one process shares one source per kept file — the watcher and the tools never renew against each other (0.831733ms)
✔ a second process finds the token the first renewed, and reuses it instead of minting over it (21.554045ms)
✔ a configured token wins, and is reported rather than minted over when the forge rejects it (2.256436ms)
✔ nothing to mint with and no token: the client says what is missing (0.167713ms)
[gitea] no tools — no Gitea token, and nothing to mint one with — set MESH_GITEA_TOKEN, or MESH_GITEA_ADMIN_USER, MESH_GITEA_ADMIN_PASSWORD_FILE, MESH_GITEA_STATE_DIR
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-qdfQzt/state/token
✔ the tools register once there is a way to a token, and the first call mints it (26.615726ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-CdErei/state/token
✔ a pull request can be closed, read and cleaned up through the tools (38.245702ms)
[gitea] no token kept — minting one
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-ZtCmZc/state/token
✔ the tools' token is minted for an account of their own, which is no administrator and writes through its team (75.224964ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-O3y1Qf/state/token
✔ with the tools' token, the forge writes a status and refuses a protection edit — the forge itself, not a guard (57.724664ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-HdA4P4/state/token
✔ what a base's protection asks is read from the branch, which the tools' account may read, not from the rules' list it may not (45.682221ms)
[gitea] the kept token was minted with other scopes than the tools take now — minting one with them
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-zbgDyX/state/token
[gitea] deleted 1 token(s) named "mesh-tools" the admin account "mesh-admin" held: the tools act as "mesh-tools" now
✔ the administrator's token from before goes: a kept admin token is minted again for the tools' account, and the forge drops the old one (43.575117ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-wqn9a1/state/token
✔ a tools' account someone made an administrator is made a plain account again at the next mint (36.502254ms)
✔ the tools never act as the administrator: MESH_GITEA_TOOLS_USER naming it is refused (0.737792ms)
✔ the manifest hands the tools their account, and the merge-ready authors name it rather than the administrator (0.863554ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-vv0HZ9/state/token
✔ a repository check said as a warning on a head without merge-check.sh reads the base's protection from the branch (40.956565ms)
[gitea] no token kept — minting one
✔ a mint that fails leaves the administrator's old token in place: the tools are never left without one (27.563481ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-D71sYj/state/token
[gitea] FAILED to delete the token(s) named "mesh-tools" the admin account "mesh-admin" holds — tried again at the next start: Gitea DELETE /users/mesh-admin/tokens/902 (named mesh-tools): 500 {"message":"the forge is not well"}
✔ an old token that cannot be deleted after the mint is said, the new token kept, and deleted at the next start (44.99018ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-pfmc9z/state/token
✔ a team another process made between the look and the create is found, not a failure (38.79625ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-5mnsNZ/state/token
✔ setting the tools' account up puts main's rule right in every repository: no push, only the tools' account merges (issue 948) (34.97649ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-XZNCUq/state/token
✔ a glob rule covering main keeps its statuses when main is given a rule of its own name (issue 948, review round 1) (37.877476ms)
✔ a start with the token kept still puts main's rules right: the module's start asks for it each time (issue 948, review round 1) (13.322083ms)
[gitea] novox/mesh-host's main moved from a2c2c35bbbbb to d28ca95ccccc: 25 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
[gitea] novox/mesh-host's main moved from d28ca95ccccc to ffffffffffff: 0 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
[gitea] novox/mesh-host: whether a trunk moved could not be weighed, so it is weighed again on the next pass — the forge is away
[gitea] novox/mesh-host's main moved from a2c2c35bbbbb to d28ca95ccccc: 0 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
[gitea] novox/mesh-controller's feat/57-config-and-provisioning moved from eeeeeeeeeeee to d28ca95ccccc: 1 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
✔ both statuses say the head and the trunk commit the verdict judged, first, whatever the summary's length (1.90469ms)
✔ a trunk's move puts an open head's merge check out of date, naming both commits, unless a check is on its way (1.126208ms)
✔ a push to a trunk puts every open pull request into it out of date — twenty-five of them — and none into another (4.201157ms)
✔ a trunk that cannot be read is weighed again on the next pass, never recorded half (0.437477ms)
✔ the trunk's move takes its turn after a verdict in flight, and weighs what the forge holds then (0.482298ms)
✔ a verdict judged against a trunk newer than the pass's first read is not put out of date (0.332956ms)
✔ a verdict that landed out of date keeps its account on the pull request (0.307235ms)
✔ a release branch's move puts every open pull request into it out of date, as a trunk's does (ADR 0358 §5) (0.558309ms)
✔ a signed event is published on the bus and acted on once (72.086334ms)
✔ an unsigned event, or one whose signature does not verify, is refused and published nowhere (41.022106ms)
✔ an event and the poll seeing the same head announce it once (0.583219ms)
✔ each forge event is named in the module's stream and trimmed (0.751332ms)
✔ the webhook is made a system webhook, which the forge sends every repository's events to (1.617907ms)
✔ default webhooks left at the receiver by an earlier start are replaced by one system webhook (1.423512ms)
✔ a repository's copy of a default webhook at the receiver is deleted, on every page, and said (5.535439ms)
✔ a forge that makes no system webhook is said, never taken as heard (0.899634ms)
✔ the receiver tells what it verified and what it refused (10.734532ms)
✔ a forge that never delivers is said loudly once, and unhealthy, when the poll finds what no event said (1.586825ms)
✔ a forge whose events stop after one delivery is said loudly once per silent spell, and refusals are unhealthy (0.62137ms)
✔ without a receiver by configuration the polls alone watch, healthy; a receiver that failed is unhealthy (0.345956ms)
✔ the polls look every 30 s until events are heard, every five minutes while they are, and 30 s again on a miss (0.322195ms)
ℹ tests 147
ℹ suites 0
ℹ pass 147
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 844.044477
**Merge check** at `b159efd6`
Judged: head `b159efd65673` merged into trunk `522780d5fcfa`; read beside it: mesh-controller `b0d489a86c23`, mesh-controller-main `b0d489a86c23`, mesh-host `cd13f7726ef0`, mesh-tools `7412c5b6c9c5`
- `mesh/merge-gate`: **PASS** — every machine composes with the change as it did without (4 of 4 compose) (modules: gitea)
- `mesh/repo-check`: **PASS** — its merge-check.sh passed, each part in its toolchain: merge-check.sh (go), merge-check-typescript.sh (typescript)
**Delivery plan** — builds gitea → novox; no bus step
- tier 0: gitea
- novox: receives gitea
- gitea provides npm-package-registry, git, gitea-read, forge-account: its consumers are sent again after it
- gitea keeps data (ADR 0233): what it holds is backed up before it moves
Run by the build seat on shanks as `build-1792193181127487471` (`builds --log build-1792193181127487471`).
```
✔ the accounts that may mark are the setting's, and only the mesh's own when it cannot be read (1.309141ms)
✔ the newest status of each context is kept, with what it said, when and by whom (20.458297ms)
✔ a commit merges only when every required status is a success; a warning blocks (0.755142ms)
✔ no rule requires nothing; a rule that cannot be read says nothing, never yes (0.306405ms)
✔ a commit is contained in another when the compare finds nothing on it the other lacks (0.947875ms)
✔ a note line is appended once, as the forge's user, in its own repository (12.451969ms)
✔ a note that cannot be written is said, never read as written (0.444877ms)
✔ one view per pull request, found by its marker; only the mesh's statuses (0.708282ms)
✔ a build machine taking the check is said on the head: building on which machine, since the local time (28.08558ms)
✔ the verdict heard while a 'building' write is in flight lands last and stays (10.853304ms)
[gitea] novox/hq@a6b08de8: the verdict of build-1791680941000000000 is an older run's than the check now pending; not written
✔ the verdict written while a start's read of the forge is in flight stays the newest status (2.976048ms)
✔ the verdict heard while a 'building' write still waits its turn: the 'building' is never written (0.672311ms)
✔ a start heard after the verdict writes nothing (0.797123ms)
✔ a start from before the check now pending (a run before a recheck, the stream's history) writes nothing (0.737832ms)
✔ an old run's verdict redelivered after a recheck's pending is not written over it (1.24762ms)
[gitea] novox/hq@a6b08de8: the verdict of build-1791680941000000000 is an older run's than the check the forge holds pending; not written
[gitea] novox/hq@a6b08de8: the verdict of build-1791680941000000000 is an older run's than the check now pending; not written
✔ an old run's verdict in flight when the recheck comes lands before the recheck's pending (4.256058ms)
✔ a rechecked head says where its new run is: building on which machine (1.056937ms)
✔ the recheck's own verdict is written, even asked a little before the pending by another machine's clock (0.805243ms)
✔ after a restart, an old run's verdict is weighed against the check the forge's history holds open (0.765172ms)
✔ a recheck the forge does not take fails its caller, and the head is as it was (0.655061ms)
✔ the status tool writes the merge check's pending through the writer, and the mesh's other statuses as they were (7.330097ms)
✔ when a check was asked is read from its id; the opening of the check now pending from the gate's history (0.272724ms)
✔ a head this process did not announce (it restarted) is said only while the forge holds its gate pending (0.368475ms)
✔ a repository as the mesh clones it is told from another (0.273354ms)
✔ gitea_api refuses the forge's administration, a branch's protection and the merge check's status, and sends nothing (5.915085ms)
✔ gitea_api refuses a file written to the default branch or one a rule guards, and writes one on a branch of its own (55.490358ms)
✔ a protection rule covers a branch as the forge decides it: classes, alternatives, a plain name in any case (issue 539) (28.805611ms)
✔ gitea_api still writes what an author writes (8.261462ms)
✔ gitea_branch_protection_set only strengthens (16.763068ms)
✔ gitea_merge_pull_request merges only a head whose required statuses all succeeded, and only that head (46.451494ms)
✔ gitea_merge_pull_request refuses a verdict judged against an earlier trunk, and asks the check once (150.821964ms)
✔ gitea_merge_pull_request merges only the head it is named, and refuses one that is not the head (17.055153ms)
✔ gitea_merge_pull_request is loud when the forge merged a head other than the one it was handed (15.068011ms)
✔ gitea_merge_pull_request says the merge happened when reading back what landed fails (26.038796ms)
✔ gitea_merge_pull_request reads back a fast-forward, and says it cannot read back a squash or a rebase (58.507426ms)
✔ gitea_delete_repo deletes only the tools' own user's repository, unguarded, untouched for 30 days, named in full (20.310085ms)
✔ gitea_api sends the path it judged, and refuses moving a pull request's base (5.279545ms)
✔ gitea_merge_pull_request merges only into a branch whose protection requires the checks (6.444373ms)
✔ a rule named for a branch is not made weaker than the glob rule that guards it now (15.122842ms)
✔ a repository owner or name is never another path (0.315465ms)
✔ the judges, without a forge (0.421917ms)
✔ main's protection is never weakened in its push or its merge whitelist (issue 948, ADR 0398 §2) (0.286835ms)
✔ gitea_api never acts as another user: sudo, in any case, and any query key it does not send are refused (3.141691ms)
✔ gitea_api never approves a pull request nor requests changes: those are the operator's (2.749674ms)
✔ no tool's caller text becomes another path or another user: `..` and `?sudo=` are refused before the forge (5.303825ms)
✔ gitea_merge_pull_request merges a verdict that read siblings only with its group passed and each sibling head reviewed (82.833885ms)
✔ a verdict put out of date by a sibling's move is refused in its own words, and asked again once (22.406428ms)
✔ every page of a pull request's files is read, though the forge caps a page at fifty (34.085295ms)
✔ a pass asks only the repositories that moved since the last look, every one before the first (0.591479ms)
✔ the directories above a merge's files that hold a module at the commit are said, the root never (14.37936ms)
✔ a lookup the forge refuses is thrown, not read as no module (5.115102ms)
✔ a renamed file is said under its old name too, as gone (novox/hq ADR 0267) (7.607402ms)
✔ a new head beyond the newest 20 open pull requests is announced, exactly once (57.942987ms)
✔ a pass pages until it reaches back past its floor, and no further (28.592758ms)
✔ gitea_branch_protection_get answers every rule, read with the administrator account the tools' token cannot read with (27.218116ms)
✔ gitea_branch_protection_get answers the rule for a branch, null when it has none (7.294806ms)
✔ gitea_branch_protection_set still gets the forge's refusal, as the tools' account (3.455335ms)
✔ the statuses asked for replace the rule's, and an empty list requires none (4.591194ms)
✔ a new rule refuses direct pushes unless asked; an administrator's override only when said (0.339095ms)
✔ main's body always refuses pushes and lets only the tools' account merge, whatever is asked (0.780973ms)
✔ another branch's body carries no merge whitelist (0.265824ms)
✔ the rule main is put right in the forge's own words: the body a setup sends carries no status checks of its own (0.333026ms)
✔ each open pull request's new head is announced once, and nothing closed or merged (24.091356ms)
✔ a verdict is the head commit's status; an error is the forge's error, never a success (1.27357ms)
✔ each layer is its own status: the gate with the modules it judged, the repository's own check beside it (0.868144ms)
✔ no verdict sets warning on a required check; only a person's decision fails (1.398602ms)
✔ a commit status is set on the commit, under the merge check's context (52.950507ms)
✔ a change plan is the gate's result: said on the status and, when it builds something, as a comment (0.648881ms)
✔ a head announced is pending in the build queue, since the local time it was announced (0.400307ms)
✔ the vault's token is used, a setting of it is ignored and said once, without its value (6.419583ms)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 at bbbbbbbbbbbb, it is now at dddddddddddd: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 was closed unmerged: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 merged onto cccccccccccc, its main is now eeeeeeeeeeee: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 was merged otherwise than its head bbbbbbbbbbbb onto cccccccccccc: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-controller#346: its verdict was judged beside mesh-host#114 was merged otherwise than its head bbbbbbbbbbbb onto cccccccccccc: out of date, nothing run (novox/hq ADR 0382)
[gitea] novox/mesh-host's main moved from 444444444444 to 333333333333: 0 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
✔ both statuses name the siblings the verdict read, keeping the verdict word and the start of its reason within 140 characters (2.529881ms)
✔ a verdict that could not learn its siblings says so on the status and in the comment, and names none (0.415717ms)
✔ a verdict without siblings is said exactly as before (an older controller) (0.817723ms)
✔ the siblings a verdict names are kept per member, replaced by its next verdict, and removed by one that read none (0.957835ms)
✔ a sibling's new head puts both statuses of the member back to pending, with the reason, and runs nothing (3.591048ms)
✔ a sibling closed unmerged puts the member's verdict out of date (0.433287ms)
✔ a move of the sibling's base puts it out of date, naming both commits (0.424007ms)
✔ the sibling's own merge of the head read onto the base read does not, in either order of the two merges (0.64603ms)
✔ a merge of another head, or onto another base, is a move (0.829733ms)
✔ only the members whose verdict names the sibling are touched, and a member that moved on is dropped (0.713082ms)
✔ a trunk that moved tells the siblings' weighing; one first seen does not (0.63278ms)
✔ the group rule names every missing thing: the group status at each head read, and each sibling head's review (1.168728ms)
✔ a merged or closed member's entry is dropped, so a sibling's later push never re-pends its merged head (0.266554ms)
✔ a consumer that asks nothing, or write, publishes: the package team (1.685737ms)
✔ a consumer that asks to read joins the readers, who write no package (0.534909ms)
✔ a reader is made, put on the readers team, and taken off the package team (0.674611ms)
✔ a publisher is made and put on the package team, as before (0.255234ms)
✔ first start: mints with the admin account, keeps the token at 0600, asks for the tools' scopes only (39.722356ms)
✔ second start: reuses the kept token, mints nothing (28.482886ms)
✔ the forge rejects the kept token (its data was restored): minted afresh, once, and the call goes through (30.112652ms)
✔ a token kept with write:admin by an earlier build is minted again without it at the next start, and the forge drops the old one (29.415911ms)
✔ a kept token lacking a scope the tools need: the forge refuses the call for the scope, the token is re-minted with the whole list, and the call goes through (29.945299ms)
✔ the kept file is gone but the forge still holds a token by that name: replaced, not refused (24.353419ms)
✔ concurrent first calls share one mint (13.424675ms)
✔ two tokens hold the name (two mints raced): both are deleted by their id, and one fresh token is minted (14.640755ms)
✔ two processes on one kept file, both without a token, mint once: the second waits for the first and reuses it (73.922913ms)
✔ a lock left by a process that died mid-mint is taken over, not waited on for ever (7.274276ms)
✔ the admin account is refused: said plainly, nothing kept, and the next call fails the same way rather than crashing (15.61371ms)
✔ one process shares one source per kept file — the watcher and the tools never renew against each other (0.831733ms)
✔ a second process finds the token the first renewed, and reuses it instead of minting over it (21.554045ms)
✔ a configured token wins, and is reported rather than minted over when the forge rejects it (2.256436ms)
✔ nothing to mint with and no token: the client says what is missing (0.167713ms)
[gitea] no tools — no Gitea token, and nothing to mint one with — set MESH_GITEA_TOKEN, or MESH_GITEA_ADMIN_USER, MESH_GITEA_ADMIN_PASSWORD_FILE, MESH_GITEA_STATE_DIR
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-qdfQzt/state/token
✔ the tools register once there is a way to a token, and the first call mints it (26.615726ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-CdErei/state/token
✔ a pull request can be closed, read and cleaned up through the tools (38.245702ms)
[gitea] no token kept — minting one
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-ZtCmZc/state/token
✔ the tools' token is minted for an account of their own, which is no administrator and writes through its team (75.224964ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-O3y1Qf/state/token
✔ with the tools' token, the forge writes a status and refuses a protection edit — the forge itself, not a guard (57.724664ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-HdA4P4/state/token
✔ what a base's protection asks is read from the branch, which the tools' account may read, not from the rules' list it may not (45.682221ms)
[gitea] the kept token was minted with other scopes than the tools take now — minting one with them
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-zbgDyX/state/token
[gitea] deleted 1 token(s) named "mesh-tools" the admin account "mesh-admin" held: the tools act as "mesh-tools" now
✔ the administrator's token from before goes: a kept admin token is minted again for the tools' account, and the forge drops the old one (43.575117ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-wqn9a1/state/token
✔ a tools' account someone made an administrator is made a plain account again at the next mint (36.502254ms)
✔ the tools never act as the administrator: MESH_GITEA_TOOLS_USER naming it is refused (0.737792ms)
✔ the manifest hands the tools their account, and the merge-ready authors name it rather than the administrator (0.863554ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-vv0HZ9/state/token
✔ a repository check said as a warning on a head without merge-check.sh reads the base's protection from the branch (40.956565ms)
[gitea] no token kept — minting one
✔ a mint that fails leaves the administrator's old token in place: the tools are never left without one (27.563481ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-D71sYj/state/token
[gitea] FAILED to delete the token(s) named "mesh-tools" the admin account "mesh-admin" holds — tried again at the next start: Gitea DELETE /users/mesh-admin/tokens/902 (named mesh-tools): 500 {"message":"the forge is not well"}
✔ an old token that cannot be deleted after the mint is said, the new token kept, and deleted at the next start (44.99018ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-pfmc9z/state/token
✔ a team another process made between the look and the create is found, not a failure (38.79625ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-5mnsNZ/state/token
✔ setting the tools' account up puts main's rule right in every repository: no push, only the tools' account merges (issue 948) (34.97649ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-tools" (write:repository, write:issue, read:user), kept at /tmp/gitea-tools-account-XZNCUq/state/token
✔ a glob rule covering main keeps its statuses when main is given a rule of its own name (issue 948, review round 1) (37.877476ms)
✔ a start with the token kept still puts main's rules right: the module's start asks for it each time (issue 948, review round 1) (13.322083ms)
[gitea] novox/mesh-host's main moved from a2c2c35bbbbb to d28ca95ccccc: 25 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
[gitea] novox/mesh-host's main moved from d28ca95ccccc to ffffffffffff: 0 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
[gitea] novox/mesh-host: whether a trunk moved could not be weighed, so it is weighed again on the next pass — the forge is away
[gitea] novox/mesh-host's main moved from a2c2c35bbbbb to d28ca95ccccc: 0 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
[gitea] novox/mesh-controller's feat/57-config-and-provisioning moved from eeeeeeeeeeee to d28ca95ccccc: 1 open pull request(s) into it put out of date, nothing run (novox/hq ADR 0323)
✔ both statuses say the head and the trunk commit the verdict judged, first, whatever the summary's length (1.90469ms)
✔ a trunk's move puts an open head's merge check out of date, naming both commits, unless a check is on its way (1.126208ms)
✔ a push to a trunk puts every open pull request into it out of date — twenty-five of them — and none into another (4.201157ms)
✔ a trunk that cannot be read is weighed again on the next pass, never recorded half (0.437477ms)
✔ the trunk's move takes its turn after a verdict in flight, and weighs what the forge holds then (0.482298ms)
✔ a verdict judged against a trunk newer than the pass's first read is not put out of date (0.332956ms)
✔ a verdict that landed out of date keeps its account on the pull request (0.307235ms)
✔ a release branch's move puts every open pull request into it out of date, as a trunk's does (ADR 0358 §5) (0.558309ms)
✔ a signed event is published on the bus and acted on once (72.086334ms)
✔ an unsigned event, or one whose signature does not verify, is refused and published nowhere (41.022106ms)
✔ an event and the poll seeing the same head announce it once (0.583219ms)
✔ each forge event is named in the module's stream and trimmed (0.751332ms)
✔ the webhook is made a system webhook, which the forge sends every repository's events to (1.617907ms)
✔ default webhooks left at the receiver by an earlier start are replaced by one system webhook (1.423512ms)
✔ a repository's copy of a default webhook at the receiver is deleted, on every page, and said (5.535439ms)
✔ a forge that makes no system webhook is said, never taken as heard (0.899634ms)
✔ the receiver tells what it verified and what it refused (10.734532ms)
✔ a forge that never delivers is said loudly once, and unhealthy, when the poll finds what no event said (1.586825ms)
✔ a forge whose events stop after one delivery is said loudly once per silent spell, and refusals are unhealthy (0.62137ms)
✔ without a receiver by configuration the polls alone watch, healthy; a receiver that failed is unhealthy (0.345956ms)
✔ the polls look every 30 s until events are heard, every five minutes while they are, and 30 s again on a miss (0.322195ms)
ℹ tests 147
ℹ suites 0
ℹ pass 147
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 844.044477
```
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Issue 955, a third fault found by #340's delivery (playbook 08). On novox the provider logged "the provider stopped: watching mesh-controller_assignments: ... reads no state ..." at 00:44:30 and never ran again. #341 fixed the grant (delivered 01:17), but only the manifest changed: the provider's bundle digest (972d97e8) did not, so the tool runner never restarted it, and no machine's key is registered.
Service.KeepRunning(service.go): main.go runssvc.KeepRunning(ctx, p.Run); whenever Run ends it starts again, pausing 5 s doubling to 2 min, said once per distinct error, reset after a 2-min run.Reviewed (larger model), approved at 4877262 (tree
5292fb9); this forge commit carries the same diff. Opened along the way: issue 988 (each restart leaves one moreboundwatch handler live: a child context per start).Delivery
novox/mesh-catalog@b159efd65673— delivered since 2026-10-16T23:32:27ZDelivery plan — builds gitea → novox; no bus step
Machines
Transitions
The commit's note under
refs/notes/mesh-plankeeps every transition:git log --notes=mesh-plan.ready at
b159efd656Reviewed by lead-57@novox: Reviewer (opus) approved 4877262 (tree
5292fb9), the same tree as this forge commit: the provider restarts a Run that ends, with backoff; restart re-takes its watches; it touches only agent- users. Marked through mesh-delivery.mark-ready (novox/hq issue 913).Merge check at
b159efd6Judged: head
b159efd65673merged into trunk522780d5fcfa; read beside it: mesh-controllerb0d489a86c23, mesh-controller-mainb0d489a86c23, mesh-hostcd13f7726ef0, mesh-tools7412c5b6c9c5mesh/merge-gate: PASS — every machine composes with the change as it did without (4 of 4 compose) (modules: gitea)mesh/repo-check: PASS — its merge-check.sh passed, each part in its toolchain: merge-check.sh (go), merge-check-typescript.sh (typescript)Delivery plan — builds gitea → novox; no bus step
Run by the build seat on shanks as
build-1792193181127487471(builds --log build-1792193181127487471).