A pending revision now holds a lost or unread message or a refusal; the watchman's finding and the view's wait say so, the take-in
event no longer requires WaitsFor (old values still decode), and the hand take-in comment loses the build.
Builds: design 47 G A-476 (b)
The watchman keeps each started step under <pipeline>/<step>/<artifact>, begins
its clock at step-started and at each mesh-wants step-progress, build seat
step-progress or step-ended of the step's ask or an ask its shared names, raises
pipeline-stands past five minutes, clears it at the step's step-ended, owes none
to a step on a person or queued for a build node, and seeds from stands at start.
Builds: design 47 G-24 §731, §732, §733 (watchman case); A-468j, A-469, A-469o, A-469p, A-469t, A-469r
mesh/delivery-group group feat/47-G-join11 rejected: composed together: the checker's verdict at the members' heads (69ffef340e92): mesh-delivery-checker
mesh/check failed (2): mesh-delivery-checker: no published of the tagged version of go-meshconditions of me…; and 1 more | 63f944b52564 on 7221a1e5cd7f
FollowSeat(mesh-wants, wants) replaces the follow by the assigner's module; the pulse and stuck keys name the holder's real
bucket taken at each use (WantsAt, AwaitsWants), a start never waiting for it and the seeded keys of any `<holder>_wants`
kept until the late read (A-321). indexer_pending and indexer_verdicts are asked of the holder of mesh-catalog on each
call. The one catch-all of the seats' events is one handler per seat (A-322), source-moved taken from the branches
follower (AdoptEvent). The manifest consumes the seat's views and the seat-holders view.
Builds: design 47 G A-210, A-318, A-321, A-322, §428
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Its finding observer, its dead-letter consumer, its reader's module and its tools' first views take the credential's module (the tools: the module the runtime serves them as) instead of constants of its name; a guard test finds none in its code.
Builds: design 47 G A-296, §428
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Its states presence, pulse and stuck and its first views were named from the constant
WatchmanModule; they are now `<cred.Module>_<state>` (watch.BucketsOf), and the watchman's
stuck keys of presence take its own presence bucket's name.
Builds: design 47 G A-296
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
signals.go BusUserWords: the node-tools special word is gone (the node-engine's node.<node> user stays, a bus account form, not a
module name). The watchman's ControlNodeModule (signals.go:591, nats) and builderModule (names.go:17) went with A-197 and WM13.
Builds: design 47 G §7, §428
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Removed with the reads of node-engine.have and mesh-host.health: the findings module-unhealthy (module <module>.<node>, token
unhealthy, urgent after four hours), kept-as-it-runs (module <module>.<node>, token kept), apply-failing from a have that failed
(machine <node>; the provisioner clients' half stays), machine-network, network-rewritten, network-unreachable, machine-units,
relogin-needed, directory-used-as-found, node-account-fallback, and the urgency of a silent control node. A want's pulse now ends on
the node-engine's running event (consumes node-engine.running).
Builds: design 47 G A-197, A-192, §541
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
record-opened is joined to the take-in.taken of its milestone, <branch>.<source revision>, whose entries name the modules; the
record-open finding is left unwritten only while one of those modules is held by an acting-part-late wait (or a store says full),
no longer while any wait is late. A record whose revision is not known taken is covered by a standing take-in-waits.
Builds: design 47 G §502, A-200
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
manifest-refused raises manifest-refused; manifest-accepted, which carries problems, raises manifest-outgrown or clears it
and any refusal; the start seeds both from indexer_verdicts. The indexer's modules view (ModuleVerdict, views.modules,
indexer_read_modules) is no longer followed, and manifest-waits is no longer written (§500).
Builds: design 47 G §500, §196, §668
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
The watchman's read of mesh-delivery-builder.builds (Build in mirror.go, the bucket, its manifest reads) goes; it hears the
builder's wait-held and wait-ended on the builder's stream, naming a package as their part, and times them as an acting
part's wait. A toolchain wait (no part) is left to the builder's own finding. The node-engine's have and mesh-host.health
reads stay: no event carries a have's phase, kept data, node record or a node's health report (asked of the lead).
Builds: design 47 G A-192, §484, §496, §602, §548
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
A hold of the composition standing past 30 minutes is delivery-waits (node, module, reason), a wait for a kept build
past 60 minutes is kept-build-waits; each clears on wait-ended. The node-engine's version is not yet named in them
(it needs the catalog entry of the fingerprint running.<node>.<node-engine's module> names).
Builds: design 47 G §499, §541, §271, §376
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
The watchman keeps the watched branch each open take-in wait names (take-in.pending's branch) and times a head only
while its branch has none; once the revision is taken in the head is timed again from its own move. The head on a
branch whose branch-unprotected stands (§497's second sentence) is not built: nothing the watchman hears says it
(the forge bridge writes that finding, no event).
Builds: design 47 G §12, §75, §497
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
A wait held with step put (no part) raises put-stands once its bound has run from its clock, and wait-ended ends it; a pipeline's
wait raises nothing for it. The seed from assigner_puts waits on that tool answering since and bound (assigner's, not built here).
Builds: design 47 G §498
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
The delivery repository's contracts and SDK live under modules/ as
mesh-delivery-contracts and mesh-delivery-sdk, their Go module paths
renamed to match the folders (the TypeScript names already did), and
every importer in mesh-delivery, mesh-catalog, mesh-host, mesh-tools
and mesh-sdk follows in the same group: a design as it should be, one
path per module, checked by the group's check building every importer.
Builds: design 47 G §427, §88
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
The watchman asks the seat that places modules who holds each build seat (seatlist.HolderOf) and watches that
holder's steps; its manifest reads the seats' steps and invokes the seat's assignments. The builder's stuck
reader looks a step up by its seat, the tests find the holders on their test seat list, the watchman's
integration bed answers who holds a seat. Its manifest tests take the manifest's requires.
Builds: design 47 G A-82, A-126
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Its handler already takes any seat's wait events; the manifest's consumes gain the two. Not run here: the watchman does not build
against mesh-contracts' meshconditionsclient as it stands (the findings writer's migration is another developer's).
Builds: design 47 G §266, §269, G-20 A-12
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
The join with CT-SD4 (6351e988) gives ParseStepKey four values (G §484, §649); the callers that read only the node or the id follow.
Builds: design 47 G §484, §649
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
Taken from 2d796fe8 (its wait half only; the store and record halves are WM7's, built). The body is the seat mesh-wants' contract Wait.
Builds: design 47 G §266, §269, §277
SeedTakeIn, run once indexer_pending answered and after the findings of a previous run were read,
ends every take-in wait, held wait and take-in-waits finding whose key the answer does not name, as a
take-in.taken would. A seed that could not ask is not run, so ends nothing.
Builds: design 47 "Taken in by the indexer it names" (A wait is bounded, said and recoverable: "At its start the watchman seeds its open waits from indexer_pending, and ends every take-in wait and take-in-waits finding it holds whose key indexer_pending no longer names"); 1966/AC195-5
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
A merge taken in with no take-in.pending said for it (read from indexer_pending at a
watchman restart, or said pending by the own-build-unknown path) ended with no taken
event, so the watchman raised take-in-waits for a wait nothing held. takeInWholeBy now
says taken for each wait WaitsOf names for the claimed entries (a revision by its oldest
merge's key, a group by its g. key) as well as for its said waits, without saying a key
twice. Drops are unchanged.
Builds: design 47 "Taken in by the indexer it names" (A wait is bounded, said and recoverable); 1966/AC195-1, AC195-2, AC195-3, AC195-4
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
The seat verb try gains members (ADR 0637 section 6); the try's group name and the missing-preview reason are one definition of mesh-delivery-sdk (ADR 0637 section 8).
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
mesh/check being carried to c78941e80b59 on 4596fbe1b7ac | from ae31b66a9e33, mesh-catalog fd4931a75261 | novox/mesh-delivery#153 in group feat/take-in
The group's check failed at 7b7096f3: the group's own manifest validator refuses the indexer's stream "take-in", outside
the bus's predefined forms. The lead's ruling: the indexer says take-in.pending and take-in.taken as its ordinary events
(its emits, mesh.mod.<module>.event.take-in.<kind>), as manifest-accepted is said, and keeps no stream. The wait's key is
in the body only: an emitted event names one event (no wildcard), and a key such as g.<group> is no event name. The
watchman consumes the two events by name on the shared events stream; it needs no replay, since it seeds its waits from
indexer_pending at its start. indexer_pending says the last event of each wait as take-in.<kind>.
Run locally, as the check runs it: cmd/mesh-judge over every manifest of mesh-catalog, mesh-delivery, mesh-host,
mesh-tools, mesh-networking and mesh-media-catalog, with this branch's and mesh-catalog's feat/take-in manifests in the
change case and origin/main's in the trunk case: 171 of 171 accepted in both. The same run with the indexer's manifest at
7b7096f3 refuses it with the check's words.
Claude-Session: https://claude.ai/code/session_01JG73KG7AfSdh6FP8VrBqCn
mesh/check failed (2): mesh-build-go: the indexer's judge at the group's mesh-delivery head 7b7096f37ea76e1…; and 1 more | 7b7096f37ea7 on ae31b66a9e33
- A take-in releases its claims (compare-and-set, only while the claim is still its own) on a lost claim, on a claim it
cannot read back, and on a merge it fails to take in: a claim left behind refused every later take-in until a
restart. T5 row with a store that fails once, both ways.
- The start step waits for nothing: with the build it runs unknown it says so and the forge consumer starts; a merged
message left unacknowledged asks the build again (clause 1). T10 row: the start answers at once and takes nothing in.
- A toolchain report evaluates pending again (ToolchainReported), since the build a revision names moves with the
toolchain. T7 row: the revision is taken in once the report makes it name the running build.
- The group's end is told by its whole subject (process.ModuleStreamSubject), so a repository named forge/group is
never taken for one; row in the indexer's handle test.
- The watchman asks the indexer's waits through indexer_pending at its start (invokes), before its durable consumer,
which does not hear again what it acknowledged before a restart; the comment that said the whole stream is heard is
corrected. W2 row: a wait seeded within its bound is raised past it.
- bridge_groups names each group's last end ask, its who, why and when, and the bridge's answer (state end-asks).
- The check's words for the indexer's rules: the trunk's manifest validator.
- The bridge's end drops the merged-state test its merge commit already implies.
- Test budget: groupask_test folded into groups_test, and the K1 mesh/check test into the owed status one: 804 unit
tests (was 806).
Claude-Session: https://claude.ai/code/session_01JG73KG7AfSdh6FP8VrBqCn
The operator, 2026-11-04: logic one module owns lives inside that module as a package of its own, so the take-in step
changes no mesh-sdk. TakeIn, its keys and event names, PendingWait and PendingAnswer move to
modules/mesh-delivery-indexer/takein (its merges a takein.Merge of its own); Forge whole, with its group fields,
ForgeGroupEnd and GroupMerge, moves to modules/mesh-delivery-forge-bridge/forgemsg. Both are nested Go modules that their
readers take by replace, as buildrun is. Every importer moves in this commit: the indexer, the bridge, the watchman and
the view. Forge's private checks (the repository name, the version) are copied small; mesh-sdk main still holds an unused
Forge until its own fold removes it.
The indexer, bridge, watchman, view and checker pin mesh-sdk 5a38a12 (origin/main), no feat/take-in commit.
The bridge's integration bus now grants the state end-asks that bridge_group_end_again added: without it every test of
the package failed on a permissions violation. Test budget: two new test functions (TestForge, TestTakeInEvents) and the
two the join added are paid by merging four pairs in internal/bridge; the repository stays at 806 unit tests.
Claude-Session: https://claude.ai/code/session_01JG73KG7AfSdh6FP8VrBqCn