indexer_take_in_pending ran the whole chain of held take-ins inside the one
loop that acts on the takeins asks, minutes long, so its answer was cut and a
second ask waited behind it and answered "holds no source-moved". The ask
now starts the chain beside the loop and answers at once, "chain started at
source <n>, <m> wait"; an ask that meets a running chain answers its state.
The chain's state (branch, source being taken in, since, how many wait) is
the key chain.<branch> of the takeins state, written at each step and
deleted at the chain's end, skipped by the asks' loop and shown by
indexer_pending. A key a stop left is deleted at the indexer's start.
Builds: design 47 H2 H-A-497 (corrected), its addendum and its second addendum
The indexer already takes a branch's held sources in order in one pass, but
nothing said so: on 22 Nov one hand ask took 114, 115 and 116 in turn, about
four minutes each, and it looked as if each needed its own hand call. Each
take-in the chain starts after one landed is now a journal line, "held
source <n> of <branch> taken in next, after <n-1>".
Builds: design 47 H2 H-A-497 (corrected)
The indexer pinned only a pipeline's own revision, so JudgePins removed
revision 292 while main.291's kept mover still had to read it. HeardPins now
pins every revision pipeline-started names in kept under the pipeline's
record, takes the latest started (said again at a re-pin) as the kept set,
freeing a kept revision it no longer names but never the pipeline's own or
its validate base, and pipeline-ended frees them all. JudgePins is
unchanged.
Builds: design 47 H2 H-A-496, H-A-496 addendum
Importers were pinned to mesh-delivery-contracts 0.1.82 to 0.1.84, reserved by
builds a competing build withdrew before release and never published, so their
builds failed "unknown revision". The series' numbered version a reservation
writes no longer pins: a pin is the series' newest released version, so a
version-reserved moves no pin and the release's re-pin moves the importers. A
pin held on a version never released is moved at the next take-in or re-pin and
said per module ("<module>: pinned to <pkg> <v>, never released; moved to <v2>").
Builds: design 47 H2 H-A-495 and its addendum (hq 9e0923ce)
Design 51 copies an upstream image by its pinned digest, and distribution's
mirror says published with no version, so nats, hello-web, mailu and showcase
could never be released. The reservation keeps its upstream refs (Upstream,
filled at the reserve from the plan's upstream tasks) and Missing never awaits
them; the indexer releases a version once every other artifact is published,
naming each upstream artifact with the pin its manifest gives, and refuses the
release loudly when an upstream artifact names no pin.
Builds: design 47 H2 H-A-486, H-A-486 (addendum)
A rerun that would reopen an ended pipeline while another open delivering
pipeline carries one of its movers at the same revision is refused with
"its pipeline <key> is running", nothing asked, so one entry is never built
side by side. Tests now cover the two paths review round 2 found untested:
the indexer keeps a revision's gathering while another pipeline's
reservation is left and removes it with the last, and a test step's rerun
reopening an ended pipeline says pipeline-started before its run is started
again.
Builds: design 47 H2 H-A-479 third addendum, H-A-479 second addendum, H-A-479 addendum (2); review round 2
A pipeline ending not passed withdrew every reservation of its revision, so
an again pipeline's failure could withdraw the head's take-in pipeline's
reservations and the other way round. The indexer now withdraws only the
reservations of the ended pipeline's own movers, read from the movers its
pipeline-ended carries, and withdraws the revision's gathering only when
none is left. A start with modules is refused for a module whose head entry
a pipeline still runs ("its pipeline <key> is running"), and takes the next
n without a bound. A rerun reopening an ended pipeline says its
pipeline-started before the step is asked again, so the revision is pinned
before anything reads it. Tests now cover the again form through the
branch's key forms and a release written into a passed again pipeline.
Builds: design 47 H2 H-A-479 second addendum, H-A-479 addendum (2); review round 1
The row proving that the judgement's strict read refuses an unknown
top-level word used class, which the H2 SDK knows, so it failed at main.
It now uses a word no SDK knows and proves the same refusal.
Builds: design 47 H2 H-A-465, H-A-469
The tests wrote registrations straight into their state or through a fake
history, so none used the stream's real subject and the one-token subject
that dropped source.novox.hq went unseen. The new integration test makes
registration-history from the manifest as the grants part does and writes a
module's and a source-only registration through the program's registry; with
the old subject it fails with the live journal's words.
Builds: design 47 H2 H-A-470 (3), H-A-408
The approved registration of source.novox.hq was given up after the bus's
last delivery without a word. On the delivery that reaches the event
deliveries bound heardDecided now gives the warrant up: its key ends refused
"not applied: <reason>" and registration-unwritten (scope module, id the
registration's name) names the ask and the reason, cleared when a later
version is written. The seed's key ends so too and says seed-waits alone.
heardDecided leaves an unapplied warrant for the bus as the other hearers do.
Builds: design 47 H2 H-A-470 (2), H-A-470 addendum (a)-(d), H-A-470 second addendum
The stream took one token after registration., so the history write of a
source-only registration (source.<owner>.<name>) got no response and the
operator's approved registration was never written. The manifest now says
registration.** (the bus's registration.>); the write subject keeps the name.
Builds: design 47 H2 H-A-470 (1), H-A-408
The running indexer read every manifest strictly for its fingerprint
inputs, so a manifest carrying a word its SDK does not know yet (class)
got no fingerprint and nothing of the revision delivered. The take-in's
read (ParseManifestForInputs) now sets an unknown top-level word aside,
as the SDK's decoder does, and takes the inputs from the words it knows;
a word inside a known one is still refused. The judgement's read
(judge.Manifest without a predefined list) stays strict.
Builds: design 47 H2 H-A-465
A module whose one process declares no bus runs that process on the
module's account, so its work-queue stream is read by exactly one process
without naming it in bus.queues; two processes, or one with a bus, still
name the reader. Lets the builder keep its process on the module's account.
Builds: design 47 H2 §793, §820; H-A-464 (ADR 0436 §6); review round 2
Review round 1 found §805's whole criterion (a release of a module
importing twenty packages of its own pipeline keeps one re-pin revision)
unit-tested only. It becomes a step of
TestASourceRevisionMovingMeshSDKNamesItAndItsImporters, the indexer whole
against an embedded bus: twenty version-reserved of mesh-sdk in main's
revision are kept in heard, its pipeline-started keeps one re-pin revision
(importer its mover, pinned to the reserved version), and the next
pipeline's reservation keeps the revision after it. The bed's indexer
durable on EVENTS now takes version-reserved, as the manifest's consumes
name it. The step fails when pipeline-started does not settle the
gathering.
Builds: design 47 H2 §805; H-A-453 (3)
Review round 1 found §802's whole criterion (a commit touching two test
files of two modules of 36 moves those two and their importers alone)
unit-tested only. It becomes a step of
TestASourceRevisionMovingMeshSDKNamesItAndItsImporters, the indexer whole
against an embedded bus: the catalogue holds 36 modules, lib-a and lib-b
among them, required by app-a (require) and app-b (require and replace);
a commit changing lib-a's and lib-b's trees alone moves lib-a, lib-b,
app-a and app-b and none of the other 32. The step fails when the
repository's commit enters the tree input, and when a derived directory
does not.
Builds: design 47 H2 §802; H-A-453 (3)
Review round 1 found §814's whole criterion (a run by hand against
seat-list gives the live validate's verdicts for the builder's and the
runner's manifests) unit-tested only. It becomes a step of
TestIndexerAnswersARunsInputOverOneMiBInChunks, the indexer started whole
against an embedded bus: with the tests' seat list in seat-definitions the
indexer keeps it in seat-lists and hands it by the verb seat-list in
chunks; the object read by its address gives the same bytes, and the
judge's verdicts for the two manifests from the verb's bytes and from a
request file read again as cmd/mesh-judge reads it are the same.
Builds: design 47 H2 §814; H-A-453 (3)
Review round 1 found the take-in's waiting bounds fixed in code
(takein_waiting.go:23-24, one and five minutes). They become the settings
take-in-waiting-line (default 1m) and take-in-waiting-bound (default 5m),
declared in the manifest and read at the start as reserved-keep is; zero
on the indexer takes the defaults. The waiting test gains a row with the
settings at 2m and 10m and checks the bounds the wait asks; the manifest
test's settings rows gain the two keys.
Builds: design 47 H2 §803; H-A-461 (2), P11
Review round 1 found heard capped at 8 MiB (module.json:137) while its
live fill is above a quarter of that. Fill read 2026-11-22 02:23 UTC with
nats.nats_buckets on novox (the bus module's show makes no entry for a
foundation part's old-section state): mesh-delivery-indexer_heard
4,654,001 bytes (4.44 MiB), 5,499 values. Four times the fill is
17.75 MiB; the smallest power of two above it is 32 MiB (H-A-437).
Builds: design 47 H2 §814, §820; H-A-461 (3), H-A-437
Review round 1 found the indexer's unit tests rose by one in H2
(seatlists_test.go:75, TestTheSeatListIsOneObjectByItsDigest). §819 as
H-A-453 (4) reads it: a module's count never rises in a change. The test
becomes the row "seat-list" of TestTheControlRecords, its steps unchanged,
so the count stands where main has it.
Builds: design 47 H2 §819, §814; H-A-453 (4)
The captured catalogue beside the judge (testdata/beside) is re-archived from
mesh-catalog 30a3e665 (feat/47-H2, the delivery group's catalogue head), so the
judge's rules are tested against manifests that name their class where they
claim a seat; the catalogue fixtures of judge and internal/index move from
835bf22b and a67f0196 to 30a3e665 (ca-trust refreshed; avahi, audit-logger and
netcheck kept at a67f0196, which they still match). Tested against mesh-sdk
a323a3fd (seatlisttest holds the group's definitions); the build seats' check
now holds current, pause and resume among the verbs the seats' definitions
serve, which have grown since 1887.
Builds: design 47 H2 §814, H-A-428, H-A-452
register takes the source form (a check of a build seat and a step kind,
no module, no path, one branch) and keeps it, its history and its
registration-changed under source.<owner>.<name>. One whose tree at the
branch's head holds a module.json at its root or under modules/, read
through the seat git's trees, is refused holds-manifest. A move of its
branch is taken in with no movers.
Builds: design 47 H2 §815, §806; H-A-349 DQ-9, H-A-438, H-A-450
A run member or mover named source.<owner>.<name> is a source-only
registration's check, which the runner runs as one step: StartRun builds
it no candidate and keeps no run, and its pipeline-ended keeps none either.
Builds: design 47 H2 §815; H-A-438, H-A-442
class core (it claims mesh-catalog). Every state names its cap. Fills read
with the bus module's nats_buckets on novox, 2026-11-22 ~02:20 CET (its
show answers only entries the bus module made, not old sections):
candidates 3,963,006 B (3.78 MiB) -> 16 MiB; manifests 5,463,959 B
(5.21 MiB) -> 32 MiB; seat-definitions 4,392,352 B (4.19 MiB) -> 32 MiB;
at 8 MiB, each at or below 2 MiB: packages 767,959; takeins 468,577;
findings 293,571; registrations 111,058; inputs 26,473; taken 5,595;
predefined 4,215; requirements 3,945; view-marks 3,068; first-views 1,554;
oldsections 256; runs 152; held 140; build-edges 0. Seventeen states, not
the sixteen the handover counted. The findings stream keeps 8 MiB, 7 days.
kept-call-pace-first 1s and kept-call-pace-max 30s, runtime settings, as
it makes kept calls.
Builds: design 47 H2 §775, §783, §810, §812, §814, §820; H-A-394 (1), H-A-411 (1) (2), H-A-437
Once the seat definitions' state is bound at the start, the indexer keeps
the seat list in seat-lists, so seat-list answers from the first delivery
and never refuses "not kept yet" for want of a trunk definition's move.
Builds: design 47 H2 §814, §820; H-A-447
The whole indexer's hand manifests of a build module and of the assigner's
holder claim a seat, so they name class core, as the judge now asks; with
none, every take-in of those tests refused them and waited out its bound.
Builds: design 47 H2 §814; H-A-394 (1)
At each trunk seat-definition move (said or not) the indexer writes the
trunk's definitions, resolved, as one JSON array sorted by seat into its
objects seat-lists under its sha256, when not held, and deletes every
object but the newest two, never the one just kept; one journal line per
write and delete. Its process serves the chunked verb seat-list: digest,
size and chunks, then the object's bytes; a list not kept yet is refused
naming it. The manifest serves seat-list and declares seat-lists in its old
objects section at 64 MiB. composer.Request gains seat-list, the digest of
the bytes Seats were filled from; the judge reads Seats alone.
Builds: design 47 H2 §814, §820; H-A-441, H-A-443 (1)
A delivery module (one claiming a seat of the delivery system, read from the
seat list's definitions) is refused for a state, log or objects entry with no
cap, in its old section or in requires; a seat-claiming manifest names class
core or leaf; finding-events is 8 MiB, 7 days; the objects template takes
max-age-days. internal/beside says a group's check reads the catalog member's
head beside it, main only when the catalog is not in the group.
Red until the captured catalog is refreshed (I6 (5)): judge
TestRulesSecrets/EveryCatalogueIdentityFitsWhatItRequires,
TestRulesSecrets/ParseManifestForInputsReadsAValueFallingBackToAPredefinedKey,
TestRulesPlacement/TheCataloguesBarPlacesBlocksAndItsShapeCoversEveryShow,
TestRulesBus/NoCatalogueManifestNamesAnInstallation; internal/index
TestAnArtifactOfNoToolchainHasNoLanguageAndNoLock,
TestAMergedManifestIsKeptOnceByItsDigestAndNamedByTheModule,
TestAModuleWithNoToolchainArtifactGetsItsFingerprintAtOnce,
TestAManifestFallingBackToAPredefinedKeyIsRead (each a captured catalog
manifest with no class, or finding-events at 16).
Builds: design 47 H2 §814, §820, §783; H-A-394 (1), H-A-411 (1) (4), H-A-415 (1), H-A-428, H-A-430 (3), H-A-434
The builder's version-reserved writes the series' numbered version with the
contents of the module's entry commit, so acceptedSeries and releasedPins pin
it; the indexer keeps each reservation in heard (reserved.<revision>.<module>,
reserved-keep) and gathers a revision's reservations until its pipeline's
start, its first step-ended or the expiry of settling.<revision>
(reserved-settle), then keeps one re-pin revision of that branch. The release's
package-published moves no fingerprint; a pipeline-ended not passed withdraws
its reservations to the newest released version.
Builds: design 47 H2 §805; H-A-418, H-A-420, H-A-424, H-A-436
The gate reads a mover's class from its catalog entry when it opens the
delivery record, so the indexer writes the manifest's class there.
Builds: design 47 H2 §810; H-A-394 (1), H-A-431 (1)
The verb registrations and the new tool indexer_registrations filter by
module and repository while they read, at most 768 KB a page, capped with
the cursor next: the tools bundle reads the state from the bus's cursor
(mesh/state.page), a store that pages none by its keys in order; total is
said when one page holds the whole read.
Builds: design 47 H2 §806, §811, §779
A branch's take-in, a run's candidate and a branch's re-pin compute as dry
runs whose packages went to the trunk's packages state. They now write them
to the state dry-runs under <branch>.<call>.packages.<name>, read through to
the live state, deleted when the dry run answers and every key at the
indexer's start; the manifests stay create-once by digest as every take-in's.
indexer_dry_runs reads the state; the manifest asks for it, 16 MiB, one
hour's TTL.
Builds: design 47 H2 §804, §820; H-A-413 and its addendum
A source-moved take-in waiting over a minute on the seat git writes one
journal line and its waiting-on (part, verb, since) in the state takeins
under running.<source revision>, removed at the take-in's end; past five
minutes it raises take-in-waiting (scope merge), cleared when the wait
ends. indexer_pending answers the running take-ins beside the held ones.
The wait is measured by a context's bound, no timer (ADR 0437 §8).
Builds: design 47 H2 §803, H-A-409
A package of another repository none publishes leaves the module without a
fingerprint (NoPackageVersion) instead of taking the gate's source head; one
of the module's own repository none publishes enters by its directory's tree,
which the tree input carries, so the same source gives the same fingerprint.
Builds: design 47 H2 §802, H-A-349 DQ-8, H-A-350 DMC-5
tags answers items {tag, commit}; trees' description says a missing directory
answers not-found; the error kind away joins refused, not-found and failed,
every answer carrying part and words with it; tag's description names the
<directory>/v<version> form. The seat's form rises to 2. The builder's,
bridge's and indexer's readers take the tag's name from the item.
Builds: design 47 H2 §776, §796; H-A-350 DMC-2, DMC-3, H-A-354
The seat generator gives every value, event and answer the wire field form
(left out while 1) and a tolerant decoder answering the form it read; each
contract carries its forms and a digest of its fields, and the compare fails
a definition whose fields changed while its form did not. Every delivery seat
is regenerated, and the readers take the form.
Builds: design 47 H2 §780; H-A-352
Review round 2. The siblings watched main, which is also the default
branch, so no test held the alike rule; they now watch stable and the
entering module is asserted to watch it.
Builds: design 47 H H-A-265, H-A-266
Review round 1. A module with a register proposal pending is skipped (the
operator's Approve registers it); a module the seed left out keeps
seed-left-out and is not registered; the branch is the first watched
branch the repository's registered modules hold alike, else its default
branch read through the seat git, and a module whose default branch is not
read is skipped and said, registered at the next kept trunk revision. One
listing of the registrations finds the modules with none, so a revision
whose modules are all registered reads no registration.
Builds: design 47 H H-A-262, H-A-263, H-A-264, H-A-265, H-A-266
Four modules added to a repository after the registrations' seed were never
registered, so a pipeline's steps for them waited on a registration nobody
asked. A kept trunk revision of any kind now registers each of its modules
with no registration, before the revision is said, with no ask: origin its
repository, path its manifest's, the first watched branch, the steps its
repository's other modules hold alike (else the seed's) and a per-artifact
status only where every sibling naming that artifact holds it identically,
said registration-changed. The start does the same once for the current
trunk graph. With no registration held the seed stays the only writer; a
module that leaves the graph keeps its registration.
Builds: design 47 H H-A-262, H-A-263, H-A-264, G A-327, ADR 0666 §2
The tool runner's approval findings of call-unanswered are of scope machine;
the indexer's entry now names scope call, so no approval key is cleared even
where an id form would match. The test's negative ids take the approval form.
Builds: design 47 H H-A-249, G §580
The daemon's call-unanswered keys from before the rename would be seeded and
stated by the tool runner's writer, which owns that kind, and never cleared.
The retired list now names call-unanswered with the indexer's own id forms
(the seed's node-build-<kind>.steps, the floor's
mesh-catalog.versions-floor.<module>, the versions fill's
mesh-delivery-builder.builder_versions), so those keys clear once and the
tool runner's approval findings stay.
Builds: design 47 H H-A-247 (3), H-A-249, G §580
The daemon's writer is handed to the process reader (event-unhandled) and the
heard writer (family-not-heard), kinds its list lacked; and call-unanswered
belongs to the tool runner's writer over the same state. The daemon's
unanswered calls are now indexer-call-unanswered (same scope and id), its list
holds the SDK parts' kinds, and merge-untaken is retired, so its live key and
condition clear at the next start. The test reads the SDK parts handed the
writer and refuses a retired kind still raised.
Builds: design 47 H H-A-242, H-A-247 (1) (2) (3), H-A-82, G §580
The daemon's writer owned every kind while the tool runner's writer shares the
state `findings`; under G §580 each bundle states only its own kinds. The
daemon's writer is now made with FindingKinds, the ten kinds its code raises,
and a unit test reads the daemon's code for every kind it gives the writer and
fails on one the list does not declare.
Builds: design 47 H H-A-240, H-A-242 (2), H-A-233, G §580
Review round 1. A branch's take-in runs through the dry indexer, which
shares the live packages state, so a branch cut before main added a
module would have deleted main's key: publish drops a kept key only on
the real indexer at the trunk's merge. And a repository whose every
asked directory is gone (one key, or all gone) returned the whole error
and raised packages-not-published at every take-in: the seat's trees of
the commit now tell a gone directory from a commit with no tree.
Builds: design 47 H H-A-232, H-A-233, H-A-234; issue 1585
A Go module mesh-catalog removed stayed in the packages state, so every
publish of mesh-catalog asked the forge for its directory, the seat git
answered not-found, and the one-directory fallback (issue 1585) ran only
on an answer without the directory: nothing of the repository was
published and its packages' digests froze unsaid.
filesVia takes the seat's not-found as a failure of the tree, so the
fallback runs; publish drops the kept key of a package whose directory
the trunk's merge no longer holds, said in the journal; a listing that
fails for another reason raises the finding packages-not-published
(scope merge, id the repository), cleared at the next publish that lists.
Builds: design 47 H H-A-232, H-A-233, H-A-234; issue 1585
The watchman's go.mod names buildrun in a replace directive with no require
line, and the walk of the module graph followed require lines only, so
buildrun's requirement of the module assigner's client was never pinned and
every test step of the watchman failed "no pin". The walk now takes the
same-repository modules a go.mod replaces by a directory as required, with
the same same-repository test, so their go.mod files are read and their
other-repository requirements pinned, transitively.
Builds: design 47 H-A-224, H-A-201
With the SDK's guard of one pull per durable (pullOnce reads the consumer's
CachedInfo and releases on the consume's Closed), the fake of a branch's
durable embedded a nil consumer and a nil consume context, so
TestABranchDurableNotThereIsMadeAndHeard panicked. The fake now answers a
ConsumerInfo with its stream, name and a created time of its own, and its
consume context closes when stopped. No other fake in mesh-delivery embeds
jetstream.Consumer; the assigner's, runner's, builder's, watchman's and
grants' unit suites pass with the new SDK unchanged.
Builds: design 47 H-A-196, H §774