Cause. mesh-delivery's recheck (checkingAgain, modules/mesh-delivery/cmd/mesh-delivery/verbs.go:291) writes pending through gitea_commit_status, which modules/gitea/tools/index.ts:643 wrote straight to the forge. The issue says the tools and the watcher share one Node process. They do not: the runtime starts each bundle the module loads as its own process (mesh-controller internal/catalogue/runtime.go, served per load; mesh-tools node-tools runtime.ts launch per entrypoint). So the write went past the watcher's GateWriter. The first run's verdict mark stayed set, so the new run never said "building on …", and an old run's verdict in flight or redelivered could land over the pending.
Change.
gitea_commit_status is now served from the watcher's bundle (index.ts), the process that holds the GateWriter. It is no longer in tools/index.ts, and its address is unchanged. A pending on mesh/merge-gate goes through the new GateWriter.rechecked. That call opens a new check of the head: it records the pending time, clears the verdict mark, and takes its turn after an old verdict in flight. A pending on mesh/repo-check takes a turn on the same chain. mesh/delivery and mesh/delivery-group are written as before. The merge check's contexts need the head's whole sha.
GateWriter.verdict takes checked.id. The ask time is read from build-<unix ns> (link.NewBuildID). A verdict whose ask predates the open check by more than the 60 s clock allowance is dropped and logged. After a restart, the open check's time is read from the forge's history of the gate (/statuses/{sha}?sort=highestindex), skipping this writer's own "building on" words. A verdict with no readable id, or a history that cannot be read, is still written.
If the recheck's write fails, the caller gets the error and the head's state is put back as it was.
#202's stand-in test (gates.pending() for a recheck) now uses the real path, gates.rechecked().
Tests. 9 new or changed tests in test/gate.test.ts. All 9 fail on main: 8 because main has no recheck path through the writer, and the restart test because the old verdict lands over the pending. The branch passes 72 of 72 in merge-check-typescript.sh for gitea. Go vet and go test -race pass for gitea and mesh-delivery, and merge-check.sh passes.
Limits. An old ask made within 60 s before the recheck is not dropped, because a lost verdict is worse than a stale one. If a recheck's pending lands but its ask then fails, the head stays pending until it is asked again. Before this change, the stale verdict could land instead.
Fixes novox/hq issue 456.
**Cause.** mesh-delivery's recheck (`checkingAgain`, modules/mesh-delivery/cmd/mesh-delivery/verbs.go:291) writes pending through `gitea_commit_status`, which modules/gitea/tools/index.ts:643 wrote straight to the forge. The issue says the tools and the watcher share one Node process. They do not: the runtime starts each bundle the module loads as its own process (mesh-controller internal/catalogue/runtime.go, `served` per load; mesh-tools node-tools runtime.ts `launch` per entrypoint). So the write went past the watcher's GateWriter. The first run's verdict mark stayed set, so the new run never said "building on …", and an old run's verdict in flight or redelivered could land over the pending.
**Change.**
- `gitea_commit_status` is now served from the watcher's bundle (index.ts), the process that holds the GateWriter. It is no longer in tools/index.ts, and its address is unchanged. A pending on mesh/merge-gate goes through the new `GateWriter.rechecked`. That call opens a new check of the head: it records the pending time, clears the verdict mark, and takes its turn after an old verdict in flight. A pending on mesh/repo-check takes a turn on the same chain. mesh/delivery and mesh/delivery-group are written as before. The merge check's contexts need the head's whole sha.
- `GateWriter.verdict` takes `checked.id`. The ask time is read from `build-<unix ns>` (link.NewBuildID). A verdict whose ask predates the open check by more than the 60 s clock allowance is dropped and logged. After a restart, the open check's time is read from the forge's history of the gate (`/statuses/{sha}?sort=highestindex`), skipping this writer's own "building on" words. A verdict with no readable id, or a history that cannot be read, is still written.
- If the recheck's write fails, the caller gets the error and the head's state is put back as it was.
- #202's stand-in test (`gates.pending()` for a recheck) now uses the real path, `gates.rechecked()`.
**Tests.** 9 new or changed tests in test/gate.test.ts. All 9 fail on main: 8 because main has no recheck path through the writer, and the restart test because the old verdict lands over the pending. The branch passes 72 of 72 in merge-check-typescript.sh for gitea. Go vet and `go test -race` pass for gitea and mesh-delivery, and merge-check.sh passes.
**Limits.** An old ask made within 60 s before the recheck is not dropped, because a lost verdict is worse than a stale one. If a recheck's pending lands but its ask then fails, the head stays pending until it is asked again. Before this change, the stale verdict could land instead.
The recheck's pending was written from the tools' bundle, its own process, past the
writer's per-head chain: the first run's verdict mark stayed set, so a rechecked head
never said where its new run builds, and an old run's verdict in flight or redelivered
could land over the pending. The status tool is now served from the watcher's bundle
and writes the merge check's pending through the writer, and a verdict whose ask
(checked.id) predates the check now pending is dropped, weighed after a restart against
the forge's own history of the gate (novox/hq issue 456).
Deliverynovox/mesh-catalog@5053e96d4e65 — delivered since 2026-10-12T10:44:21Z
Delivery plan — builds gitea → novox; no bus step
tier 0: gitea
novox: receives gitea
gitea provides npm-package-registry, git: its consumers are sent again after it
gitea keeps data (ADR 0233): what it holds is backed up before it moves
Machines
build-agent on ace: passed — carried by plan-1791800868526452268
route-proxy on ace: passed — carried by plan-1791800868526452268
gitea on novox: passed — healthy 1 times over 15s
gitea on the rest running it: sent
Transitions
2026-10-12 10:20 (new) → proposed (announced): novox/mesh-catalog#234's head announced
2026-10-12 10:24 proposed → checked (checked): the verdict names this commit
2026-10-12 10:24 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
2026-10-12 10:26 ready → proposed (recheck): approved; main moved (#233 added the gitea forge bundle) after its checks ran; judge #234 again before merge (issue 456)
2026-10-12 10:28 proposed → checked (checked): the verdict names this commit
2026-10-12 10:28 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
2026-10-12 10:41 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 10:41 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
2026-10-12 10:44 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@5053e96d4e65` — **delivered** since 2026-10-12T10:44:21Z
**Delivery plan** — builds gitea → novox; no bus step
- tier 0: gitea
- novox: receives gitea
- gitea provides npm-package-registry, git: its consumers are sent again after it
- gitea keeps data (ADR 0233): what it holds is backed up before it moves
**Machines**
- build-agent on ace: passed — carried by plan-1791800868526452268
- route-proxy on ace: passed — carried by plan-1791800868526452268
- gitea on novox: passed — healthy 1 times over 15s
- gitea on the rest running it: sent
**Transitions**
- 2026-10-12 10:20 (new) → proposed (announced): novox/mesh-catalog#234's head announced
- 2026-10-12 10:24 proposed → checked (checked): the verdict names this commit
- 2026-10-12 10:24 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
- 2026-10-12 10:26 ready → proposed (recheck): approved; main moved (#233 added the gitea forge bundle) after its checks ran; judge #234 again before merge (issue 456)
- 2026-10-12 10:28 proposed → checked (checked): the verdict names this commit
- 2026-10-12 10:28 checked → ready (accepted): the gate passed or warned, and the repository's own check did not fail
- 2026-10-12 10:41 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 10:41 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
- 2026-10-12 10:44 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/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: 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 novox as build-1791800441430938361 (builds --log build-1791800441430938361).
--- the gate: gitea
gitea: ok, keeps data (valuable); WARNING: server stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 0
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T10:14:58Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T10:14:58Z
modules the change moves:
- gitea (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 1 module(s) in 1 tier(s): gitea
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.007s
words: 12 retired words and 70 daemons' lines, none in the 42350 strings of 682 files
ok git.novox.be/novox/mesh-catalog/checks/agenthome 0.012s
agent-home: no archive and no file written into under the agent's home
ok git.novox.be/novox/mesh-catalog/checks/paths 0.007s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 69, as kept
testing modules/gitea
ok gitea/cmd/npm-registry 1.860s
--- its merge-check-typescript.sh (typescript toolchain)
type-checking modules/gitea (TypeScript)
testing modules/gitea (TypeScript)
✔ a pull request opened after the margin is looked at, by its own update time (45.103538ms)
✔ every open pull request is read with what its base requires and what its head has (gitea_open_pulls) (158.783962ms)
✔ a pull request or issue number that is not one is refused before the forge is asked (1.834813ms)
✔ the newest status of each context is kept, with what it said, when and by whom (27.064411ms)
✔ a commit merges only when every required status is a success; a warning blocks (1.698963ms)
✔ no rule requires nothing; a rule that cannot be read says nothing, never yes (0.589521ms)
✔ a commit is contained in another when the compare finds nothing on it the other lacks (1.199992ms)
✔ a note line is appended once, as the forge's user, in its own repository (39.59309ms)
✔ a note that cannot be written is said, never read as written (0.794981ms)
✔ one view per pull request, found by its marker; only the mesh's statuses (1.135892ms)
✔ a build machine taking the check is said on the head: building on which machine, since the local time (53.099021ms)
✔ the verdict heard while a 'building' write is in flight lands last and stays (8.868723ms)
[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 (1.201151ms)
✔ the verdict heard while a 'building' write still waits its turn: the 'building' is never written (0.981001ms)
✔ a start heard after the verdict writes nothing (1.331042ms)
✔ a start from before the check now pending (a run before a recheck, the stream's history) writes nothing (0.583451ms)
✔ an old run's verdict redelivered after a recheck's pending is not written over it (1.552313ms)
[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 (4.999897ms)
✔ a rechecked head says where its new run is: building on which machine (0.900502ms)
✔ the recheck's own verdict is written, even asked a little before the pending by another machine's clock (0.669271ms)
✔ after a restart, an old run's verdict is weighed against the check the forge's history holds open (0.821011ms)
✔ a recheck the forge does not take fails its caller, and the head is as it was (0.58696ms)
✔ the status tool writes the merge check's pending through the writer, and the mesh's other statuses as they were (14.154571ms)
✔ when a check was asked is read from its id; the opening of the check now pending from the gate's history (0.272301ms)
✔ a head this process did not announce (it restarted) is said only while the forge holds its gate pending (2.505023ms)
✔ a repository as the mesh clones it is told from another (1.672722ms)
✔ gitea_api refuses the forge's administration, a branch's protection and the merge check's status, and sends nothing (6.044769ms)
✔ gitea_api refuses a file written to the default branch or one a rule guards, and writes one on a branch of its own (87.470694ms)
✔ gitea_api still writes what an author writes (12.663079ms)
✔ gitea_branch_protection_set only strengthens (20.537112ms)
✔ gitea_merge_pull_request merges only a head whose required statuses all succeeded, and only that head (39.89218ms)
✔ gitea_merge_pull_request merges only the head it is named, and refuses one that is not the head (15.856544ms)
✔ gitea_merge_pull_request is loud when the forge merged a head other than the one it was handed (9.928495ms)
✔ gitea_merge_pull_request says the merge happened when reading back what landed fails (21.010072ms)
✔ gitea_merge_pull_request reads back a fast-forward, and says it cannot read back a squash or a rebase (46.898652ms)
✔ gitea_delete_repo deletes only the tools' own user's repository, unguarded, untouched for 30 days, named in full (23.034425ms)
✔ gitea_api sends the path it judged, and refuses moving a pull request's base (4.676137ms)
✔ gitea_merge_pull_request merges only into a branch whose protection requires the checks (6.48057ms)
✔ a rule named for a branch is not made weaker than the glob rule that guards it now (15.374454ms)
✔ a repository owner or name is never another path (0.24865ms)
✔ the judges, without a forge (0.307601ms)
✔ gitea_api never acts as another user: sudo, in any case, and any query key it does not send are refused (2.275133ms)
✔ gitea_api never approves a pull request nor requests changes: those are the operator's (3.194295ms)
✔ no tool's caller text becomes another path or another user: `..` and `?sudo=` are refused before the forge (5.354928ms)
✔ every page of a pull request's files is read, though the forge caps a page at fifty (59.692091ms)
✔ a pass asks only the repositories that moved since the last look, every one before the first (1.551862ms)
✔ the directories above a merge's files that hold a module at the commit are said, the root never (21.138192ms)
✔ a lookup the forge refuses is thrown, not read as no module (6.59992ms)
✔ a renamed file is said under its old name too, as gone (novox/hq ADR 0267) (9.617175ms)
✔ the statuses asked for replace the rule's, and an empty list requires none (9.658994ms)
✔ a new rule refuses direct pushes unless asked; an administrator's override only when said (0.39264ms)
✔ each open pull request's new head is announced once, and nothing closed or merged (19.926231ms)
✔ a verdict is the head commit's status; an error is the forge's error, never a success (4.016096ms)
✔ each layer is its own status: the gate with the modules it judged, the repository's own check beside it (0.819322ms)
✔ no verdict sets warning on a required check; only a person's decision fails (1.093161ms)
✔ a commit status is set on the commit, under the merge check's context (116.487377ms)
✔ a change plan is the gate's result: said on the status and, when it builds something, as a comment (0.754591ms)
✔ a head announced is pending in the build queue, since the local time it was announced (0.309941ms)
✔ first start: mints with the admin account, keeps the token at 0600, asks for the tools' scopes only (91.142239ms)
✔ second start: reuses the kept token, mints nothing (23.652086ms)
✔ the forge rejects the kept token (its data was restored): minted afresh, once, and the call goes through (47.863823ms)
✔ 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 (21.339102ms)
✔ 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 (17.291976ms)
✔ the kept file is gone but the forge still holds a token by that name: replaced, not refused (16.766996ms)
✔ concurrent first calls share one mint (9.931765ms)
✔ the admin account is refused: said plainly, nothing kept, and the next call fails the same way rather than crashing (11.399967ms)
✔ one process shares one source per kept file — the watcher and the tools never renew against each other (0.918692ms)
✔ a second process finds the token the first renewed, and reuses it instead of minting over it (17.949837ms)
✔ a configured token wins, and is reported rather than minted over when the forge rejects it (2.274163ms)
✔ nothing to mint with and no token: the client says what is missing (0.18568ms)
[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-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-56qtYj/state/token
✔ the tools register once there is a way to a token, and the first call mints it (8.138763ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-jyP2UB/state/token
✔ a pull request can be closed, read and cleaned up through the tools (22.906105ms)
ℹ tests 72
ℹ suites 0
ℹ pass 72
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 633.038766
**Merge check** at `5053e96d`
- `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: 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 novox as `build-1791800441430938361` (`builds --log build-1791800441430938361`).
```
--- the gate: gitea
gitea: ok, keeps data (valuable); WARNING: server stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 0
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T10:14:58Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T10:14:58Z
modules the change moves:
- gitea (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes
- a machine (yvbpsm): composes
a merge would rebuild 1 module(s) in 1 tier(s): gitea
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.007s
words: 12 retired words and 70 daemons' lines, none in the 42350 strings of 682 files
ok git.novox.be/novox/mesh-catalog/checks/agenthome 0.012s
agent-home: no archive and no file written into under the agent's home
ok git.novox.be/novox/mesh-catalog/checks/paths 0.007s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 69, as kept
testing modules/gitea
ok gitea/cmd/npm-registry 1.860s
--- its merge-check-typescript.sh (typescript toolchain)
type-checking modules/gitea (TypeScript)
testing modules/gitea (TypeScript)
✔ a pull request opened after the margin is looked at, by its own update time (45.103538ms)
✔ every open pull request is read with what its base requires and what its head has (gitea_open_pulls) (158.783962ms)
✔ a pull request or issue number that is not one is refused before the forge is asked (1.834813ms)
✔ the newest status of each context is kept, with what it said, when and by whom (27.064411ms)
✔ a commit merges only when every required status is a success; a warning blocks (1.698963ms)
✔ no rule requires nothing; a rule that cannot be read says nothing, never yes (0.589521ms)
✔ a commit is contained in another when the compare finds nothing on it the other lacks (1.199992ms)
✔ a note line is appended once, as the forge's user, in its own repository (39.59309ms)
✔ a note that cannot be written is said, never read as written (0.794981ms)
✔ one view per pull request, found by its marker; only the mesh's statuses (1.135892ms)
✔ a build machine taking the check is said on the head: building on which machine, since the local time (53.099021ms)
✔ the verdict heard while a 'building' write is in flight lands last and stays (8.868723ms)
[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 (1.201151ms)
✔ the verdict heard while a 'building' write still waits its turn: the 'building' is never written (0.981001ms)
✔ a start heard after the verdict writes nothing (1.331042ms)
✔ a start from before the check now pending (a run before a recheck, the stream's history) writes nothing (0.583451ms)
✔ an old run's verdict redelivered after a recheck's pending is not written over it (1.552313ms)
[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 (4.999897ms)
✔ a rechecked head says where its new run is: building on which machine (0.900502ms)
✔ the recheck's own verdict is written, even asked a little before the pending by another machine's clock (0.669271ms)
✔ after a restart, an old run's verdict is weighed against the check the forge's history holds open (0.821011ms)
✔ a recheck the forge does not take fails its caller, and the head is as it was (0.58696ms)
✔ the status tool writes the merge check's pending through the writer, and the mesh's other statuses as they were (14.154571ms)
✔ when a check was asked is read from its id; the opening of the check now pending from the gate's history (0.272301ms)
✔ a head this process did not announce (it restarted) is said only while the forge holds its gate pending (2.505023ms)
✔ a repository as the mesh clones it is told from another (1.672722ms)
✔ gitea_api refuses the forge's administration, a branch's protection and the merge check's status, and sends nothing (6.044769ms)
✔ gitea_api refuses a file written to the default branch or one a rule guards, and writes one on a branch of its own (87.470694ms)
✔ gitea_api still writes what an author writes (12.663079ms)
✔ gitea_branch_protection_set only strengthens (20.537112ms)
✔ gitea_merge_pull_request merges only a head whose required statuses all succeeded, and only that head (39.89218ms)
✔ gitea_merge_pull_request merges only the head it is named, and refuses one that is not the head (15.856544ms)
✔ gitea_merge_pull_request is loud when the forge merged a head other than the one it was handed (9.928495ms)
✔ gitea_merge_pull_request says the merge happened when reading back what landed fails (21.010072ms)
✔ gitea_merge_pull_request reads back a fast-forward, and says it cannot read back a squash or a rebase (46.898652ms)
✔ gitea_delete_repo deletes only the tools' own user's repository, unguarded, untouched for 30 days, named in full (23.034425ms)
✔ gitea_api sends the path it judged, and refuses moving a pull request's base (4.676137ms)
✔ gitea_merge_pull_request merges only into a branch whose protection requires the checks (6.48057ms)
✔ a rule named for a branch is not made weaker than the glob rule that guards it now (15.374454ms)
✔ a repository owner or name is never another path (0.24865ms)
✔ the judges, without a forge (0.307601ms)
✔ gitea_api never acts as another user: sudo, in any case, and any query key it does not send are refused (2.275133ms)
✔ gitea_api never approves a pull request nor requests changes: those are the operator's (3.194295ms)
✔ no tool's caller text becomes another path or another user: `..` and `?sudo=` are refused before the forge (5.354928ms)
✔ every page of a pull request's files is read, though the forge caps a page at fifty (59.692091ms)
✔ a pass asks only the repositories that moved since the last look, every one before the first (1.551862ms)
✔ the directories above a merge's files that hold a module at the commit are said, the root never (21.138192ms)
✔ a lookup the forge refuses is thrown, not read as no module (6.59992ms)
✔ a renamed file is said under its old name too, as gone (novox/hq ADR 0267) (9.617175ms)
✔ the statuses asked for replace the rule's, and an empty list requires none (9.658994ms)
✔ a new rule refuses direct pushes unless asked; an administrator's override only when said (0.39264ms)
✔ each open pull request's new head is announced once, and nothing closed or merged (19.926231ms)
✔ a verdict is the head commit's status; an error is the forge's error, never a success (4.016096ms)
✔ each layer is its own status: the gate with the modules it judged, the repository's own check beside it (0.819322ms)
✔ no verdict sets warning on a required check; only a person's decision fails (1.093161ms)
✔ a commit status is set on the commit, under the merge check's context (116.487377ms)
✔ a change plan is the gate's result: said on the status and, when it builds something, as a comment (0.754591ms)
✔ a head announced is pending in the build queue, since the local time it was announced (0.309941ms)
✔ first start: mints with the admin account, keeps the token at 0600, asks for the tools' scopes only (91.142239ms)
✔ second start: reuses the kept token, mints nothing (23.652086ms)
✔ the forge rejects the kept token (its data was restored): minted afresh, once, and the call goes through (47.863823ms)
✔ 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 (21.339102ms)
✔ 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 (17.291976ms)
✔ the kept file is gone but the forge still holds a token by that name: replaced, not refused (16.766996ms)
✔ concurrent first calls share one mint (9.931765ms)
✔ the admin account is refused: said plainly, nothing kept, and the next call fails the same way rather than crashing (11.399967ms)
✔ one process shares one source per kept file — the watcher and the tools never renew against each other (0.918692ms)
✔ a second process finds the token the first renewed, and reuses it instead of minting over it (17.949837ms)
✔ a configured token wins, and is reported rather than minted over when the forge rejects it (2.274163ms)
✔ nothing to mint with and no token: the client says what is missing (0.18568ms)
[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-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-56qtYj/state/token
✔ the tools register once there is a way to a token, and the first call mints it (8.138763ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-jyP2UB/state/token
✔ a pull request can be closed, read and cleaned up through the tools (22.906105ms)
ℹ tests 72
ℹ suites 0
ℹ pass 72
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 633.038766
```
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: 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 ace as build-1791800804691400885 (builds --log build-1791800804691400885).
--- the gate: gitea
gitea: ok, keeps data (valuable); WARNING: server stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 0
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T10:23:58Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T10:23:58Z
modules the change moves:
- gitea (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes; the change adds 0 resource(s) and removes 1 (-archive:gitea.bundle-forge)
- a machine (yvbpsm): composes
a merge would rebuild 1 module(s) in 1 tier(s): gitea
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.003s
words: 12 retired words and 70 daemons' lines, none in the 42350 strings of 682 files
ok git.novox.be/novox/mesh-catalog/checks/agenthome 0.014s
agent-home: no archive and no file written into under the agent's home
ok git.novox.be/novox/mesh-catalog/checks/paths 0.006s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 69, as kept
testing modules/gitea
ok gitea/cmd/npm-registry 1.794s
--- its merge-check-typescript.sh (typescript toolchain)
type-checking modules/gitea (TypeScript)
testing modules/gitea (TypeScript)
✔ a pull request opened after the margin is looked at, by its own update time (7.273385ms)
✔ every open pull request is read with what its base requires and what its head has (gitea_open_pulls) (41.886314ms)
✔ a pull request or issue number that is not one is refused before the forge is asked (1.302971ms)
✔ the newest status of each context is kept, with what it said, when and by whom (20.943153ms)
✔ a commit merges only when every required status is a success; a warning blocks (0.577837ms)
✔ no rule requires nothing; a rule that cannot be read says nothing, never yes (0.216177ms)
✔ a commit is contained in another when the compare finds nothing on it the other lacks (0.677234ms)
✔ a note line is appended once, as the forge's user, in its own repository (10.130169ms)
✔ a note that cannot be written is said, never read as written (0.418257ms)
✔ one view per pull request, found by its marker; only the mesh's statuses (0.603635ms)
✔ a build machine taking the check is said on the head: building on which machine, since the local time (13.982957ms)
✔ the verdict heard while a 'building' write is in flight lands last and stays (7.386237ms)
[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 (1.386388ms)
✔ the verdict heard while a 'building' write still waits its turn: the 'building' is never written (0.584229ms)
✔ a start heard after the verdict writes nothing (0.616731ms)
✔ a start from before the check now pending (a run before a recheck, the stream's history) writes nothing (0.646135ms)
✔ an old run's verdict redelivered after a recheck's pending is not written over it (0.67496ms)
[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 (1.767485ms)
✔ a rechecked head says where its new run is: building on which machine (0.664129ms)
✔ the recheck's own verdict is written, even asked a little before the pending by another machine's clock (0.550315ms)
✔ after a restart, an old run's verdict is weighed against the check the forge's history holds open (1.125999ms)
✔ a recheck the forge does not take fails its caller, and the head is as it was (0.404501ms)
✔ the status tool writes the merge check's pending through the writer, and the mesh's other statuses as they were (5.333405ms)
✔ when a check was asked is read from its id; the opening of the check now pending from the gate's history (0.174799ms)
✔ a head this process did not announce (it restarted) is said only while the forge holds its gate pending (0.432323ms)
✔ a repository as the mesh clones it is told from another (0.199495ms)
✔ gitea_api refuses the forge's administration, a branch's protection and the merge check's status, and sends nothing (4.31072ms)
✔ gitea_api refuses a file written to the default branch or one a rule guards, and writes one on a branch of its own (47.896742ms)
✔ gitea_api still writes what an author writes (7.570513ms)
✔ gitea_branch_protection_set only strengthens (18.778419ms)
✔ gitea_merge_pull_request merges only a head whose required statuses all succeeded, and only that head (29.507195ms)
✔ gitea_merge_pull_request merges only the head it is named, and refuses one that is not the head (15.887311ms)
✔ gitea_merge_pull_request is loud when the forge merged a head other than the one it was handed (11.915617ms)
✔ gitea_merge_pull_request says the merge happened when reading back what landed fails (22.689458ms)
✔ gitea_merge_pull_request reads back a fast-forward, and says it cannot read back a squash or a rebase (47.664386ms)
✔ gitea_delete_repo deletes only the tools' own user's repository, unguarded, untouched for 30 days, named in full (22.1852ms)
✔ gitea_api sends the path it judged, and refuses moving a pull request's base (3.641361ms)
✔ gitea_merge_pull_request merges only into a branch whose protection requires the checks (7.403308ms)
✔ a rule named for a branch is not made weaker than the glob rule that guards it now (13.93616ms)
✔ a repository owner or name is never another path (0.296979ms)
✔ the judges, without a forge (0.394012ms)
✔ gitea_api never acts as another user: sudo, in any case, and any query key it does not send are refused (5.018371ms)
✔ gitea_api never approves a pull request nor requests changes: those are the operator's (2.507417ms)
✔ no tool's caller text becomes another path or another user: `..` and `?sudo=` are refused before the forge (5.051534ms)
✔ every page of a pull request's files is read, though the forge caps a page at fifty (26.233706ms)
✔ a pass asks only the repositories that moved since the last look, every one before the first (0.352684ms)
✔ the directories above a merge's files that hold a module at the commit are said, the root never (12.369262ms)
✔ a lookup the forge refuses is thrown, not read as no module (4.385231ms)
✔ a renamed file is said under its old name too, as gone (novox/hq ADR 0267) (4.748785ms)
✔ the statuses asked for replace the rule's, and an empty list requires none (5.855015ms)
✔ a new rule refuses direct pushes unless asked; an administrator's override only when said (0.252034ms)
✔ each open pull request's new head is announced once, and nothing closed or merged (7.416672ms)
✔ a verdict is the head commit's status; an error is the forge's error, never a success (0.744521ms)
✔ each layer is its own status: the gate with the modules it judged, the repository's own check beside it (0.579199ms)
✔ no verdict sets warning on a required check; only a person's decision fails (1.008367ms)
✔ a commit status is set on the commit, under the merge check's context (31.434832ms)
✔ a change plan is the gate's result: said on the status and, when it builds something, as a comment (0.433676ms)
✔ a head announced is pending in the build queue, since the local time it was announced (0.234892ms)
✔ first start: mints with the admin account, keeps the token at 0600, asks for the tools' scopes only (31.610171ms)
✔ second start: reuses the kept token, mints nothing (16.937105ms)
✔ the forge rejects the kept token (its data was restored): minted afresh, once, and the call goes through (19.533221ms)
✔ 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 (21.94641ms)
✔ 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 (17.693408ms)
✔ the kept file is gone but the forge still holds a token by that name: replaced, not refused (17.622284ms)
✔ concurrent first calls share one mint (10.973867ms)
✔ the admin account is refused: said plainly, nothing kept, and the next call fails the same way rather than crashing (12.204061ms)
✔ one process shares one source per kept file — the watcher and the tools never renew against each other (1.143561ms)
✔ a second process finds the token the first renewed, and reuses it instead of minting over it (20.392558ms)
✔ a configured token wins, and is reported rather than minted over when the forge rejects it (2.484785ms)
✔ nothing to mint with and no token: the client says what is missing (0.2601ms)
[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-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-m2sXiZ/state/token
✔ the tools register once there is a way to a token, and the first call mints it (8.642779ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-BcBQhf/state/token
✔ a pull request can be closed, read and cleaned up through the tools (22.304183ms)
ℹ tests 72
ℹ suites 0
ℹ pass 72
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 425.615611
**Merge check** at `5053e96d`
- `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: 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 ace as `build-1791800804691400885` (`builds --log build-1791800804691400885`).
```
--- the gate: gitea
gitea: ok, keeps data (valuable); WARNING: server stay(s) up and say(s) not how it is ready — judged by liveness alone, refused from 2026-11-18 (ADR 0240 rule 8)
long-running resources without health: 1
files asking for a setting without saying whether it is trusted: 0
1 manifest(s) checked. Judged against the seats this binary carries; a claim on one of the mesh's own seats is judged fully at registration, and a seat declared by a module not given here reads as unknown. Identities judged on a 6-character machine name, against the bounds of the providers given here
judging 1 module(s) changed against 4 machine(s), as the snapshot of 2026-10-12T10:23:58Z says they run
composing every machine as the mesh is
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
composing every machine with the change
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
the bus's user list leaves out 5 user(s) the mesh has minted no credential for: controller, node.f35, node.idt, node.ygcyr, node.yvbpsm. Each is a user that cannot connect until one is issued
merge gate: PASS — every machine composes with the change as it did without (4 of 4 compose)
judged by controller development build against the facts of 2026-10-12T10:23:58Z
modules the change moves:
- gitea (changed)
machines:
- a machine (f35): composes
- holds mesh-dns-resolver (idt): composes
- holds git, holds mesh-artifact-store, holds mesh-broker, holds mesh-catalog, holds mesh-delivery, holds mesh-dns-resolver, holds mesh-store, holds mesh-vault, holds npm-package-registry, the control node, the hub (ygcyr): composes; the change adds 0 resource(s) and removes 1 (-archive:gitea.bundle-forge)
- a machine (yvbpsm): composes
a merge would rebuild 1 module(s) in 1 tier(s): gitea
--- its merge-check.sh (go toolchain)
ok git.novox.be/novox/mesh-catalog/checks/words 0.003s
words: 12 retired words and 70 daemons' lines, none in the 42350 strings of 682 files
ok git.novox.be/novox/mesh-catalog/checks/agenthome 0.014s
agent-home: no archive and no file written into under the agent's home
ok git.novox.be/novox/mesh-catalog/checks/paths 0.006s
paths: no module names another's state on its machine (3 known, listed to go)
long-running resources without health: 69, as kept
testing modules/gitea
ok gitea/cmd/npm-registry 1.794s
--- its merge-check-typescript.sh (typescript toolchain)
type-checking modules/gitea (TypeScript)
testing modules/gitea (TypeScript)
✔ a pull request opened after the margin is looked at, by its own update time (7.273385ms)
✔ every open pull request is read with what its base requires and what its head has (gitea_open_pulls) (41.886314ms)
✔ a pull request or issue number that is not one is refused before the forge is asked (1.302971ms)
✔ the newest status of each context is kept, with what it said, when and by whom (20.943153ms)
✔ a commit merges only when every required status is a success; a warning blocks (0.577837ms)
✔ no rule requires nothing; a rule that cannot be read says nothing, never yes (0.216177ms)
✔ a commit is contained in another when the compare finds nothing on it the other lacks (0.677234ms)
✔ a note line is appended once, as the forge's user, in its own repository (10.130169ms)
✔ a note that cannot be written is said, never read as written (0.418257ms)
✔ one view per pull request, found by its marker; only the mesh's statuses (0.603635ms)
✔ a build machine taking the check is said on the head: building on which machine, since the local time (13.982957ms)
✔ the verdict heard while a 'building' write is in flight lands last and stays (7.386237ms)
[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 (1.386388ms)
✔ the verdict heard while a 'building' write still waits its turn: the 'building' is never written (0.584229ms)
✔ a start heard after the verdict writes nothing (0.616731ms)
✔ a start from before the check now pending (a run before a recheck, the stream's history) writes nothing (0.646135ms)
✔ an old run's verdict redelivered after a recheck's pending is not written over it (0.67496ms)
[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 (1.767485ms)
✔ a rechecked head says where its new run is: building on which machine (0.664129ms)
✔ the recheck's own verdict is written, even asked a little before the pending by another machine's clock (0.550315ms)
✔ after a restart, an old run's verdict is weighed against the check the forge's history holds open (1.125999ms)
✔ a recheck the forge does not take fails its caller, and the head is as it was (0.404501ms)
✔ the status tool writes the merge check's pending through the writer, and the mesh's other statuses as they were (5.333405ms)
✔ when a check was asked is read from its id; the opening of the check now pending from the gate's history (0.174799ms)
✔ a head this process did not announce (it restarted) is said only while the forge holds its gate pending (0.432323ms)
✔ a repository as the mesh clones it is told from another (0.199495ms)
✔ gitea_api refuses the forge's administration, a branch's protection and the merge check's status, and sends nothing (4.31072ms)
✔ gitea_api refuses a file written to the default branch or one a rule guards, and writes one on a branch of its own (47.896742ms)
✔ gitea_api still writes what an author writes (7.570513ms)
✔ gitea_branch_protection_set only strengthens (18.778419ms)
✔ gitea_merge_pull_request merges only a head whose required statuses all succeeded, and only that head (29.507195ms)
✔ gitea_merge_pull_request merges only the head it is named, and refuses one that is not the head (15.887311ms)
✔ gitea_merge_pull_request is loud when the forge merged a head other than the one it was handed (11.915617ms)
✔ gitea_merge_pull_request says the merge happened when reading back what landed fails (22.689458ms)
✔ gitea_merge_pull_request reads back a fast-forward, and says it cannot read back a squash or a rebase (47.664386ms)
✔ gitea_delete_repo deletes only the tools' own user's repository, unguarded, untouched for 30 days, named in full (22.1852ms)
✔ gitea_api sends the path it judged, and refuses moving a pull request's base (3.641361ms)
✔ gitea_merge_pull_request merges only into a branch whose protection requires the checks (7.403308ms)
✔ a rule named for a branch is not made weaker than the glob rule that guards it now (13.93616ms)
✔ a repository owner or name is never another path (0.296979ms)
✔ the judges, without a forge (0.394012ms)
✔ gitea_api never acts as another user: sudo, in any case, and any query key it does not send are refused (5.018371ms)
✔ gitea_api never approves a pull request nor requests changes: those are the operator's (2.507417ms)
✔ no tool's caller text becomes another path or another user: `..` and `?sudo=` are refused before the forge (5.051534ms)
✔ every page of a pull request's files is read, though the forge caps a page at fifty (26.233706ms)
✔ a pass asks only the repositories that moved since the last look, every one before the first (0.352684ms)
✔ the directories above a merge's files that hold a module at the commit are said, the root never (12.369262ms)
✔ a lookup the forge refuses is thrown, not read as no module (4.385231ms)
✔ a renamed file is said under its old name too, as gone (novox/hq ADR 0267) (4.748785ms)
✔ the statuses asked for replace the rule's, and an empty list requires none (5.855015ms)
✔ a new rule refuses direct pushes unless asked; an administrator's override only when said (0.252034ms)
✔ each open pull request's new head is announced once, and nothing closed or merged (7.416672ms)
✔ a verdict is the head commit's status; an error is the forge's error, never a success (0.744521ms)
✔ each layer is its own status: the gate with the modules it judged, the repository's own check beside it (0.579199ms)
✔ no verdict sets warning on a required check; only a person's decision fails (1.008367ms)
✔ a commit status is set on the commit, under the merge check's context (31.434832ms)
✔ a change plan is the gate's result: said on the status and, when it builds something, as a comment (0.433676ms)
✔ a head announced is pending in the build queue, since the local time it was announced (0.234892ms)
✔ first start: mints with the admin account, keeps the token at 0600, asks for the tools' scopes only (31.610171ms)
✔ second start: reuses the kept token, mints nothing (16.937105ms)
✔ the forge rejects the kept token (its data was restored): minted afresh, once, and the call goes through (19.533221ms)
✔ 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 (21.94641ms)
✔ 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 (17.693408ms)
✔ the kept file is gone but the forge still holds a token by that name: replaced, not refused (17.622284ms)
✔ concurrent first calls share one mint (10.973867ms)
✔ the admin account is refused: said plainly, nothing kept, and the next call fails the same way rather than crashing (12.204061ms)
✔ one process shares one source per kept file — the watcher and the tools never renew against each other (1.143561ms)
✔ a second process finds the token the first renewed, and reuses it instead of minting over it (20.392558ms)
✔ a configured token wins, and is reported rather than minted over when the forge rejects it (2.484785ms)
✔ nothing to mint with and no token: the client says what is missing (0.2601ms)
[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-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-m2sXiZ/state/token
✔ the tools register once there is a way to a token, and the first call mints it (8.642779ms)
[gitea] no token kept — minting one
[gitea] the forge already holds a token named "mesh-tools" — replacing it
[gitea] minted a token for "mesh-admin" (write:repository, write:issue, read:user), kept at /tmp/gitea-BcBQhf/state/token
✔ a pull request can be closed, read and cleaned up through the tools (22.304183ms)
ℹ tests 72
ℹ suites 0
ℹ pass 72
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 425.615611
```
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.
Fixes novox/hq issue 456.
Cause. mesh-delivery's recheck (
checkingAgain, modules/mesh-delivery/cmd/mesh-delivery/verbs.go:291) writes pending throughgitea_commit_status, which modules/gitea/tools/index.ts:643 wrote straight to the forge. The issue says the tools and the watcher share one Node process. They do not: the runtime starts each bundle the module loads as its own process (mesh-controller internal/catalogue/runtime.go,servedper load; mesh-tools node-tools runtime.tslaunchper entrypoint). So the write went past the watcher's GateWriter. The first run's verdict mark stayed set, so the new run never said "building on …", and an old run's verdict in flight or redelivered could land over the pending.Change.
gitea_commit_statusis now served from the watcher's bundle (index.ts), the process that holds the GateWriter. It is no longer in tools/index.ts, and its address is unchanged. A pending on mesh/merge-gate goes through the newGateWriter.rechecked. That call opens a new check of the head: it records the pending time, clears the verdict mark, and takes its turn after an old verdict in flight. A pending on mesh/repo-check takes a turn on the same chain. mesh/delivery and mesh/delivery-group are written as before. The merge check's contexts need the head's whole sha.GateWriter.verdicttakeschecked.id. The ask time is read frombuild-<unix ns>(link.NewBuildID). A verdict whose ask predates the open check by more than the 60 s clock allowance is dropped and logged. After a restart, the open check's time is read from the forge's history of the gate (/statuses/{sha}?sort=highestindex), skipping this writer's own "building on" words. A verdict with no readable id, or a history that cannot be read, is still written.gates.pending()for a recheck) now uses the real path,gates.rechecked().Tests. 9 new or changed tests in test/gate.test.ts. All 9 fail on main: 8 because main has no recheck path through the writer, and the restart test because the old verdict lands over the pending. The branch passes 72 of 72 in merge-check-typescript.sh for gitea. Go vet and
go test -racepass for gitea and mesh-delivery, and merge-check.sh passes.Limits. An old ask made within 60 s before the recheck is not dropped, because a lost verdict is worse than a stale one. If a recheck's pending lands but its ask then fails, the head stays pending until it is asked again. Before this change, the stale verdict could land instead.
Delivery
novox/mesh-catalog@5053e96d4e65— delivered since 2026-10-12T10:44:21ZDelivery 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.Merge check at
5053e96dmesh/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 novox as
build-1791800441430938361(builds --log build-1791800441430938361).Merge check at
5053e96dmesh/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 ace as
build-1791800804691400885(builds --log build-1791800804691400885).