The pin judge removed the branch's previous revision in the very write that took the next one in
(pipelines main.13 and main.15 on 2026-11-20), so the validate step's case before could not be read
and the composition was never asked. The take-in now pins the newest revision it kept before its
write under the same pipeline record as its own, and the pipeline's end frees the record on every
key it pins; a take-in with no revision before it pins only its own and says so.
Builds: H-A-147 (hq design/47-H1-answers-59 310f8955ba16a1ca8ae9661d8196cab16e3d9597), design 47 H §743, §749, H-A-108 (3)
As the grants part gives a module's consumer on a seat event's stream kept by age:
without a start from now a made durable would replay every past source-moved of the
branch past the start's guard; the current revision is read at the start already.
Builds: design 47 H §743, H-A-134 (hq 8d30cc42), H-A-129, H-A-130
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
The grants part composes no per-branch durable on the source-moved stream, so the
indexer never heard a source revision after its start. Its bind now makes the
durable itself, as a module's seat-event consumer is made (SeatEventConsumerConfig,
named <node>_<indexer>_<branch>, filtered to the branch, one message at a time),
and binds an existing one as it stands.
Builds: design 47 H §743, H-A-134
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
At a clean state the grants part makes the per-branch consumers of source-moved only from the modules view, which the
indexer answers only after its first take-in, so nothing ever took the trunk in. At its start, and again at each
failed bind, the indexer now reads the branch's current source revision through mesh-delivery.source and takes it in
as a heard source-moved of it would, when its view keeps no revision of the branch at or above it: start true exactly
when it keeps none, every repository the answer names a head. No timer; binding on registration-changed is unchanged.
Builds: design 47 H §743, §768, H-A-129, H-A-130
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
A source-moved the indexer cannot take in is written into the state held, the event whole
(call, heads, start beside branch, source, why, since), and acknowledged once written; a later
source-moved of a held branch is held behind it, and the branch's key keeps them in order.
The indexer takes them in again from that state alone, in their order, on its own room and
at its start; one neither taken in nor written stays unacknowledged. The whole test's held
case reads the new form and that nothing is left unacknowledged.
Builds: design 47 H §744, H-A-114, H-A-118
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
The seat mesh-delivery-checker's generated Go seat, client and TypeScript package go from
mesh-delivery-contracts (and their entries in its manifest), with the old copies under
contracts/go; the indexer, the forge bridge and the assigner require them no more, and the
assigner judges with its own internal/assignment verdict constants. The git, node-build-go
and node-build-node fixtures, the buildstep and steprun tests and the indexer's and
assigner's whole tests name the seat mesh-wants (the asker of every check-kind run in H1)
or the runner where they named the checker; the indexer's unused CheckerModule and the
judge's stream exemption for the checker go. The step kinds check and recheck stay with
their queues (H-A-117).
Builds: design 47 H H-A-57, H-A-117
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
Nothing of the one model reads which build the indexer runs: the take-in holds nothing for the
indexer a revision names and indexer_pending answers the state held alone, so FixOwnBuild, the
state own-build, the ask of the node-engine's composition and its invoke go. No pull request's
candidate is written any more, so what only it read goes: the member heads' packages
(candidatePublished), SameCandidate, headOf and CandidatesKey; and consumerSink with the state
consumers (H-A-80). The state taken stays: the view reads it. The run on request's candidate
stays (H-A-105 b). The manifest test holds the state and the invoke absent.
Builds: design 47 H §743 (holds nothing for the indexer), §764 (the indexer); H-A-48, H-A-80, H-A-102, H-A-105
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
The whole test starts the indexer against an embedded bus with the gate's source-moved on its
kept stream and a durable per watched branch: a first revision with start true and no movers;
one revision per take-in carrying its movers, pinned pipeline.<branch>.<source> and said as
take-in.taken; the same source-moved given again said from the kept movers with no second
revision; pipeline-started pinning, a restart keeping a running pipeline's revision; a take-in
on a full store held, said once as take-in.pending and in the state held, taken again on the
indexer's own room once pipeline-ended frees a pin; another branch's held take-in leaving
main going on, skipped by hand with why in the hand-act log; a branch watched later bound on
registration-changed. The packages test takes a mesh-sdk commit in and names mesh-sdk and the
module requiring it, not the one requiring nothing of it. The toolchain test drives take-ins;
the tests of the forge messages, the verdicts, the pins events of the gate and the view moving
between take-ins (H-A-49, H-A-103) go. 7 integration tests (was 13).
Builds: design 47 H §743, §744, §749, §768; H-A-12, H-A-30, H-A-49, H-A-50, H-A-90, H-A-92, H-A-97, H-A-102, H-A-103
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
TakeIn, TakeInRepository, the catch-up, IndexRegistered and RetryUnread go: a module is taken in
only from a source revision, and the gate's gate_head_take is the one hand take of a head. The
manifest serves no indexer_take_in and names no catch-up tool beside the catalog revisions.
Builds: design 47 H §742, §764 (the indexer); H-A-48, H-A-105 (a)
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
keepPull, dropPull, ApplyVerdict and the pull form of the seat-definitions key go with the
pull and open messages and the checker's verdict; indexer_seat_definitions_again and
indexer_seat_definition name the trunk alone.
Builds: design 47 H §764 (the indexer); H-A-48
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
No ledger writes the state consumers any more, so the tool indexer_consumer, its reader and
the state's declaration go; the manifest test holds their absence.
Builds: design 47 H §764 (the indexer); H-A-80
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
I6: indexer_unpin is renamed indexer_pin_drop and takes the pin's key <branch>.<revision>
and why: it drops every record's pin of that key, clears its finding pin-missing and the
removal rule then applies; a hand act recorded in the hand-act log before it is asked
(§765, H-A-57).
I5 follow-up: the asks of indexer_take_in_pending and indexer_take_in_skip were refused by
the ask reader (DecodeTakeIn named no held branch); they are read now, and
indexer_pending answers {held: [{branch, source, why, since}]}, since in RFC 3339, the
state held's form (H-A-102). The test drives each tool's ask through the reader.
Builds: design 47 H §765 (indexer_pin_drop), §744 (the held tools); H-A-57, H-A-97, H-A-102
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
A source-moved the indexer cannot take in (the catalog store full, a read unanswered) stays
unacknowledged on its branch's consumer, so that branch's next waits while another branch's
go on; it is written in the new state held, one key per branch {branch, source, why, since}
(H-A-97), and said once per source as take-in.pending {branch, source, why}. A full store
sets heldFull, so RoomAgain says room once it has room, and the indexer's own room (H-A-50),
heard on its events consumer, takes every held one again, never a timer. indexer_pending
reads the state held; indexer_take_in_pending takes a branch's held one again now and
indexer_take_in_skip, with branch, source and why, acknowledges it unwritten, says
take-in.taken with no entries, skipped, by and why, and is recorded in the hand-act log;
both are asks in the state takeins. The held key goes at the take-in or the skip.
Builds: design 47 H §744; H-A-12, H-A-50, H-A-97
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
The indexer binds one durable consumer per watched branch on the stream of the gate's
source-moved, named <node>_<indexer>_<escaped branch> through mesh-sdk's
Account.NamedEventConsumer; it takes each message in with TakeInSource and acknowledges it
once its revision and take-in.taken are written, and leaves it unacknowledged when the
take-in fails. The watched branches are the registrations'; a newly watched branch's
consumer is bound on the indexer's own registration-changed, which the manifest now
consumes. A source-moved of the old form (no call) is refused for good with a journal line.
Builds: design 47 H §743 (one durable per watched branch), §765 (the old form refused); H-A-30, H-A-90
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
TakeInSource takes a gate's source-moved in: the trees of each repository it moves at the
named commit through the seat git, the fingerprints of the modules registered for it, the
movers against the branch's newest catalog revision, and one revision <escaped
branch>.<source> for every branch, the trunk included, numbered by the source and written
once by read before write. The revision carries its movers, and a redelivery whose revision
is kept says take-in.taken again from them, recomputing nothing (H-A-92). It is pinned for
the take-in's pipeline as pipeline.<branch>.<source> in the same write when it has a mover,
none for a take-in with no mover. take-in.taken carries branch, source, revision, call and
the movers' entries alone, under a duplicate id of its branch and source (H-A-31). A
source-moved with no call is refused as the old form (§765).
Only a take-in keeps a catalog revision (H-A-49): the recompute, the build ends' imports and
the start's reconcile keep none, and KeepView, keepEntries, keepTrunkRevision and
PinPipelines go. Another branch than the first is taken in on its newest revision's view.
Builds: design 47 H §743, §749 (the pin in the take-in's write), §765 (the old form); H-A-31, H-A-32, H-A-49, H-A-92
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
I8: the take-in fed by the forge bridge's messages goes with B1's forgemsg: the FORGE
consumer and its handlers, the group end, the pending state with its keys, claims, dry
take-in and hand tools (indexer_pending, indexer_take_in_pending, indexer_drop_pending,
indexer_take_in_merge, kept again for the held take-in in I5), the catch-up by comparison
(kickBehind, TakeInBehind, sayUntaken, takeInHeads, kickCatchUp, CatchUp, merge-untaken),
the source numbering in memory (nextSource), the source verb's follow and Src stamping on
a source answer, the bridge's merged message call, the pull request candidates written
from pull and open messages, the checker's previews and verdict hearing. A repository's
trees at a commit are the internal whole (wholeAt), which ApplyMessage takes in; the dry
indexer stays for a run on request's candidate. The manifest drops the consumes of
forge.**, the checker's verdict and version-delivered, the reads of the checker's
previews and the invokes of open-pulls, bridge_merged_message and mesh-delivery.source.
RoomAgain says room when a take-in is held on a full store (heldFull, I5 holds it).
H-A-82: the fill of delivered versions into the edges goes: HearVersionDelivered, the
delivered.<module> key, the first fill's kinds 2, 4 and 6 from the placement and the
stream running, and outdated's kinds 2, 4 and 6; their entries read not yet delivered.
The tests of the removed paths go (index_test's candidate groups, source_test,
branch_take_in_test, previews_test, candidate_manifests_test, the FORGE handle tests);
the versions tests now hold kinds 2, 4 and 6 as declarations. The integration tests
still drive FORGE and are rewritten in I10.
Builds: design 47 H §743, §744, §764 (the indexer); H-A-48, H-A-82
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
I7: the indexer judges no manifest. The judge pass, its verdicts and unjudged states,
the events manifest-accepted and manifest-refused, the tools verdicts, indexer_verdicts,
indexer_judge and indexer_say_accepted, and the take-in's wait for the verdicts state
go; indexer_bus_state answers from the modules state at its mark. The seed writes
validate required in every policy. oldsections stays (H-A-79).
I6, the half that follows C-R: the gate's pinned, freed, heads-replaced and
pins-stated are no longer heard. A pipeline-started pins the revision it names under
pipeline.<branch>.<source>, its pipeline-ended frees it, each on a durable of its own
stream; kept revisions are judged against the pins the state holds, never a count
(made and pins-stated go). A take-in pins its revision once, for its one pipeline.
indexer_pin_drop (§765) is still to do.
The entry's pull goes (§763). The tests of the old take-in (pending, behind, pushed)
go with it, ahead of I8. The unit suite is green except manifest_test's check that
the watchman's and the grants part's invokes of the indexer's verdicts tools are
served: their removal is the contracts developer's (H-A-79).
Builds: design 47 H §743, §745, §749, §763, §764 (the indexer); H-A-48, H-A-49, H-A-79
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
Every account may call every verb: the verb revision answers every caller, so its
callers list and the node-engine account's mapping go; the first-view tools take no
callers list and the tools bundle's env carries none; the seat-list part is made
without one.
Red only from the contract removals C1 and C3 (I7's manifest events, the gate's pin
events, the entry's pull), followed in the next commits.
Builds: design 47 H §752, §764 (every seat contract: callers); H-A-41, H-A-66
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
The mesh knows no pull request (design 47 G, G-1 §38-39), so the indexer keeps no pull request's config declarations.
Removed: the code that writes the state (control_records.go: the candidate writes after a merge and for a pull
message), its bucket, view family (views.config-candidates) and read tool (indexer_read_config_candidates) in the
indexer's module.json, the view's tool indexer_config_candidates and its read of the indexer's tool, the state and its
fixtures in the seat mesh-catalog's definition (generated Go and TypeScript regenerated by seatgen). indexer_config_readers
no longer names an open pull request as a reader of a key. The state `predefined` stays.
Builds: ADR 0681 §2 (the indexer keeps no config candidate: state, read tool and view, the seat contract's layout; `predefined` stays), §1 (indexer_config_readers names no pull request).
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
FillFloor clears the module's call-unanswered finding and its call-since key on every exit but the end of its context (a release
gave the module a version, it left the trunk, the callee answered with a failure), as ADR 0668 §3 has: FloorModules never
picks the module again. The module's clock starts at its first unanswered try, not when the form gives up. The integration test
sets the form's bound to one second, waits for the module's call-since key in the state heard, and only then serves git.tags;
it fails with `return err` at the unanswered try. Unit cases for the three exits.
Builds: design 47 G §412, A-490g, A-490h, A-490i
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
A-490i corrects A-490h: the growing-pause timer is gone (floorpause.go and the guard's exemption removed; ADR 0437 §8 stands).
Each module's floor is read on its own (FloorModules, FillFloor); a read is a kept call through mesh-sdk's form, asked again
at once when the form gives up, so it has no bound; the first unanswered try starts the module's clock and the finding
call-unanswered is written five minutes later and cleared on answer. The integration test leaves the forge unserved until
a delay, and the fill follows with no restart and no revision.
Builds: design 47 G §412, A-490c, A-490d, A-490e, A-490f, A-490g, A-490h, A-490i
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
The floor of a module with no version is the highest among its Go tags and the versions each package holder answers through
its seat's verb versions (npm-package-registry, go-package-registry; the sdk's npm 0.1.18 above its Go tags 0.1.16). An
unanswered read is asked again on a growing pause (5 s, 15 s, 30 s, 1, 2, 5 min) until it answers, with call-unanswered five
minutes after the first. A floor entry holds no artifacts, so a package-published heard later at its version is recorded as
its release and says version-released. One whole-indexer integration test on a real bus with a stand-in seat git and holder.
Builds: design 47 G §412, A-490c, A-490d, A-490e, A-490f, A-490g, A-490h
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
mesh-catalog.versions answered [] for mesh-sdk although the forge holds go/v0.1.12..v0.1.16: the first fill read only
builder_versions, so a module the builder held nothing of was filled with nothing, and the once-only key closed it. The
indexer now reads, for every module of the newest trunk revision the versions hold none of, every tag of its Go package
from the seat git's verb tags (repository of its catalog entry, prefix the package's directory + "/v", "v" at the root),
fills the highest as its version, at each start and each catalog revision kept, and raises call-unanswered after five
minutes when the forge does not answer.
Builds: design 47 G §412, A-490b, A-490c, A-490d, A-490e
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
KeptTrunk also holds the newest revision and its entries, and the manifests the entries carry by digest.
The verb revision asked of the trunk's newest and the manifests the verbs module and provides read come
from it, with no read of the bucket or the manifests state; a candidate, a branch and an older number
still read the bucket. A failed re-read says the view is BEHIND and the next heard catalog-revision
reads again. The follower's state is named catalog-half.
Builds: design 47 G §421, lead answers A-487, A-487b, A-487c (03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md)
KeptTrunk reads the newest trunk revision from catalog-revisions once at the start (a follower of the
bundle's views.Reader, adopted on the seat's event catalog-revision) and re-reads once for each heard
event of a newer trunk revision; reads list nothing. The invoke seat:mesh-catalog.revision and the
meshcatalogclient follower are removed; the consume stays.
Builds: design 47 G §421, lead answers A-487 and A-487b (03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md)
The tools bundle's view of the trunk (KeptTrunk) is made from the half of the catalog its own follower
of the seat holds (meshcatalogclient.WireCatalog on a views.Reader: the newest revision once at the
start, again on the seat's event catalog-revision), once per revision, and says each read with its
revision, module count and time. A read no longer lists catalog-revisions or fetches the newest
revision object. The verb module reads the one module it is asked (CatalogRead.Module). The manifest
consumes mesh-catalog.catalog-revision and invokes seat:mesh-catalog.revision for the follower.
Builds: design 47 G §421, lead answer A-487 (03-DESIGN/01-to-be/47-milestones/G-22-lead-answers-2.md)
The tools bundle's verbs module, modules, provides and dependents read the modules and the module graph key by key from
the newest trunk revision, and each key listed the whole catalog-revisions bucket again (6 s per call at 11 000 objects,
35 s with 12 callers); seat-definitions read its state once per seat. The view is read once per call (SourceValues), the
seat definitions come from a watched copy, a manifest is read once by its digest.
Builds: design 47 G (the seat mesh-catalog's verbs served by the indexer's tools bundle; ADR 0604 §2, ADR 0580 clause 1)
The recompute of a merge leaves a module of another repository as held unless its imported package versions changed
(ADR 0593 §3 stands). TestIndexerCountsOnlyThePlacedBuildNodesToolchains reports language versions (A-471).
Builds: design 47 G A-476a, A-471
(a) validation_wait.go and validate.go go with the validation ask, its end consumer, indexer_validate_again and
Pending.Validation; (b) evaluate always takes the revision in with the build the program runs (holdFor and the build
comparison go); (c) a merge's recompute writes only the modules registered for its repository.
Builds: design 47 G A-476 (a), (b), (c)
The two package-published of mesh-delivery-pipeline 0.1.6 sit on SEAT_mesh-builds_package-published, where the
indexer's program durable novox_mesh-delivery-indexer holds them pending and was never pulled: the program bound only
its durable on EVENTS, and a kept seat event is never there. So no version was released. The same holds for
toolchain-agreed (2 pending) and node-build-go's step-ended (10 pending). The program now binds, once its EVENTS
durable is bound, its durable on each kept stream whose event it handles, as the builder does.
Builds: design 47 G A-397, §349
The indexer leaves an event it cannot take for the bus to give again, and the bus drops it after the fifth delivery
its durable allows without either saying so: a package-published lost so leaves a version unreleased silently. On
that last delivery the indexer now says it gives the message up, naming its subject, what it carries (for a
package-published the artifact and version) and the last refusal.
Builds: design 47 G A-420
The bus refused the bare subject mesh.seat.mesh-catalog.event.catalog-revision: the seat definition
(mesh-delivery-contracts/seats/mesh-catalog/seat.json) gives the event the subject catalog-revision.<branch>, and the
holder is granted catalog-revision.*. The publish now uses meshcatalog.SubjectCatalogRevision(the event's branch); an event
with no branch is refused loudly, never given a bare subject.
Builds: design 47 G, the seat definition mesh-catalog (event catalog-revision, subject catalog-revision.<branch>)
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
lead answer A-391: the start replay keeps the view once (measured: 50 min CPU per start).
applyBuildEnd reports whether it wrote; ApplyBuildEnd keeps the view only then;
ApplyBuildEndsAtStart applies every entry and keeps the view once if any wrote.
Builds: design 47 G lead answer A-391
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
The grants part follows the first view modules on the indexer's views family after its first ask. G's indexer keeps
no state modules, so nothing said its changes and a newly merged module reached grants only when it asked again. A
new trunk revision that changes the modules is now said as the mark of modules at that revision's number (the number
the first view answers every value at), after which a reader short of it asks the first view once; views.modules is
among the indexer's emits.
Builds: design 47 G A-360, A-354
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
index.Module is gone: the indexer and its registry carry the credential's module (findings' observer and ids, its own
entry and graph key, the history subject, the views emitter and reader), the tools take the served-as module.
Builds: design 47 G A-296, §428
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
buildrun's AssignerHolder asked the seat's verb for the holder every 5 seconds. The builder and the checker now follow
mesh-module-assigner's assignments through views.ProcessReader.FollowSeat with the catalog half's holder function (the
checker gains the catalog half's reader of §614 for it); the indexer reads the claimant from its own catalog
(Indexer.ClaimantNow) and follows its seat-holders view again at each refresh of its build modules. AssignerHolder goes; the
manifests consume the seat-holders view (the checker also the catalog half's events and verbs), and the assigner's
seat-holders first-view tool answers the checker and the indexer. The beds give each part a claimant and a seat-holders view
in place of the seat verb's answer; the indexer's beds hold the claimant from the start, so the merge test reads the newest
revision.
Builds: design 47 G §717, §614, A-298, A-302
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
Its 26 `<Module>_<state>` bucket constants (candidates, manifests, verdicts, packages, taken, the
control records, pins, pending and the rest), its consumers, findings and first views were named
from the constant Module; they are now index.Bucket(cred.Module, <state>), and its words name
the state where they named the bucket.
Builds: design 47 G A-296
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
A-163: the asker in `accept.<queue>.<asker>` is the asking seat's name (mesh-pipeline-runner,
mesh-catalog, mesh-wants, mesh-builds), never a nickname: the builder asks as mesh-builds (was
"builder"), the indexer's validation as mesh-catalog (was "indexer"), the assigner's runs as
mesh-wants (was "assigner"), the runner's steps on `accept.check.mesh-pipeline-runner` and
`accept.delivery.mesh-pipeline-runner` (were `accept.check.step` and `accept.delivery.step`, where
the holder read `step` as the asker). An asker naming none (the checker) asks on main's subject
(A-184). A build seat's queues are delivery, check and recheck alone (A-170, A-184): buildstep's
AskKinds holds no step, a holder takes a step from check or delivery, and the grants part makes a
worker per queue, never per kind. A holder keys a named ask's step `<asker>.<node>.<id>` and main's
`<node>.<id>` (§481, §483); readers watch their own name's keys. The composition step's step-progress
names the one caller mesh-wants' definition lists for step, read from the seat list (A-294).
Builds: design 47 G §481, §483, §543, §600, A-163, A-170, A-184, A-248, A-294
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h