Commit Graph
1087 Commits
Author SHA1 Message Date
jschoubben 25b3030a93 S2 take once in a bundle too, Go and TypeScript
A bundle holder takes an ask once through the runtime's create with the
ask's bound as the key's time to live: stdio.KeptState.TakeOnce and
TypeScript takeOnce; stdio.TakenState and TakenKey are the one name, which
process refers to.

Builds: design 47 H2 §775, §820; H-A-359
2026-11-21 23:12:58 +01:00
jschoubben 423de0e8a2 S3 away: read a holder's answer of the kind away as a transport fault
A holder answers error-kind away, with part and words, when a part it
depends on is away; the kept call reads it from the answer and calls again
at the pace. Go: busretry.ErrorKindAway, AwayAnswer, AwayIn, Away;
TypeScript: ERROR_KIND_AWAY, awayAnswer, awayIn.

Builds: design 47 H2 §776, §775; H-A-349 DQ-15, H-A-350 DMC-2, H-A-354
2026-11-21 23:10:27 +01:00
jschoubben d40fa0b3a0 Carry a process's drain bound in the declaration
A process that stays up may name drain-bound (seconds): the node-engine
writes its unit's stop from it, so the tool runner finishes its calls in
flight before it exits. Absent is none; a step or a scheduled run is
refused one.

Builds: design 47 H2 §813 (12); H-A-358, H-A-353
2026-11-21 23:08:38 +01:00
jschoubben f5dc3b6bd1 S2 take once: keep the ids a holder took in its state taken
A holder of a kept call takes an ask once by its id: TakeOnce creates
<seat>.<ask> in the state taken with the ask's bound as its time to live,
and a second ask of the same id within it answers false.

Builds: design 47 H2 §775, §820; H-A-338 (6), H-A-339 (4)
2026-11-21 23:08:38 +01:00
jschoubben 14c2532b0c S1 kept call paced: carry the ask id, pace at the settings' defaults with jitter, stop at the caller's bound
The kept call is made again on a transport fault after kept-call-pace-first
(1 s) doubling to kept-call-pace-max (30 s), each wait within 20 %, until
the caller's bound; past it the answer names every call's time and words.
One journal line per seat at the first fault, then at most once a minute.
The ask id travels in the header x-ask; a holder reads it with
process.AskOf, a bundle with stdio.AskIDOf (TypeScript askIdOf) from the
call's _meta mesh/ask.

Builds: design 47 H2 §775; H-A-351
2026-11-21 23:06:24 +01:00
jschoubben fedf9d2e7b Pull small batches in the event hub and keep slow handlers alive (H-A-281)
The hub pulled the bus client's default 500 messages per pull and handled
them one by one, so buffered messages outlived the 30 s ack wait and were
given again and handled twice (met live on the builder: 37 of 38
outstanding redelivered). HearEvents, the one pull of the hub and of
HearStream, now pulls at most twenty messages per pull and says in
progress for a message whose handler runs past a third of the consumer's
ack wait, read once from the bound consumer's information.

Builds: design 47 H H-A-281
2026-11-21 16:48:05 +01:00
mesh-tools dfcf2147bc Merge pull request 'Design 47 H1, proof 2: one comment in the SDK, delivered with its importer in one merge call' (#206) from proof/47-H1-sdk into main 2026-11-21 11:27:05 +00:00
jschoubben 78807e5ef3 Design 47 H1: a held writer always starts; a retired entry may name its scope
Start returned at the first early write, early clear or retired clear that
failed, before the writer started, so one refused kind stopped the writer
whole. A refused kind is now skipped (refuse said it once) and any other
failure is said in the journal and left to the next start; the writer always
starts. A RetiredKind may name its Scope, so an entry clears only keys of it.

Builds: design 47 H H-A-242 (3), H-A-249, G §580
2026-11-21 11:47:20 +01:00
jschoubben d06835bf93 Design 47 H1: a retired kind may name the ids its writer once wrote
Where another writer of the same state owns a retired kind now, clearing every
key of it would clear that writer's findings. A RetiredKind entry names its
kind and, optionally, the exact ids and id prefixes this writer wrote under
it; Start clears only those keys, every key of the kind when none are named.

Builds: design 47 H H-A-247 (3), H-A-249, G §580
2026-11-21 11:42:45 +01:00
jschoubben 220d511f95 Design 47 H1: a held writer clears its retired kinds once at start and reads only its own keys
A kind no code raises any more stood in the state with its condition open and
nobody to clear it. A writer's Retired kinds are cleared once at Start, said
as finding-cleared so the keeper clears their conditions, and never seeded;
Start reads (Get) only its own and its retired keys, checking ownership first.

Builds: design 47 H H-A-242, H-A-247 (3), G §580
2026-11-21 11:39:08 +01:00
jschoubben af2abcf121 Design 47 H1: a held writer seeds only the kinds it owns; an unowned kind is said once
The tool runner's held writer seeded every key of the module's shared state
`findings`, the daemon's kinds included, and then refused them once a minute,
so no finding of the indexer was stated. Start now seeds only keys of the
writer's own kinds (G §580: another bundle states its own), and a kind the
writer does not own is refused with "finding kind X not declared", said in
the journal once per kind and left out of a statement, never retried.

Builds: design 47 H H-A-240, H-A-242 (1) (3), G §580
2026-11-21 11:28:04 +01:00
jschoubben 3bc7da52e4 Design 47 H1, proof 2: mark the SDK side of the two-module change
The process package's doc names the proof, so the SDK publishes a new version that node-tools, its importer, is built against in the same merge call.

Builds: design 47 H §5 step 7, G §370
2026-11-21 01:46:34 +01:00
jschoubben 4a5eb26f4e views: read the durable's ack floor up to a mark, not its pending count, after the handlers spoke
The views test read NumAckPending as soon as the picture and the adopted handler had said the last event, but the
reader acknowledges after both, and the bus takes acks on its own; on the loaded build node (pipeline main.54) one
ack was still pending. The reader binds the module's durable once (the hub's one pull); the guard of H-A-196 is not
involved. The test now publishes a mark the hub hands after every event before it, acknowledged with the bus's
answer, and reads that the ack floor reached the mark; an event the owner says after it may still wait.

Builds: design 47 H §774, H-A-196, H-A-197
2026-11-21 01:20:35 +01:00
jschoubben becaf8096a H-A-196: one SDK function claims, pulls and wraps a durable's loop
The tool runner would otherwise copy the claim's wrapper (review round 1, P7):
ConsumeOnce claims the durable, pulls it and answers the loop, and HearEvents
uses it; Conn.Consume claims through the same guard. A drain keeps its claim
until the loop ended, and the loop's Closed is closed only once the claim is
released, so a bind during a drain is refused and one after it is never.

Builds: design 47 H H-A-196, H-A-197
2026-11-20 22:18:43 +01:00
jschoubben c1c369f50c H-A-196: a process pulls a durable consumer once
A second pull loop on a durable the same process already pulls takes every other
event from the first, and nothing said so: the tool runner's two loops on the
indexer bundle's durable left the catalog seat's revision verb stale twice today.
The SDK now refuses it at bind time, naming the durable and both binders:
PullOnce claims a durable (by stream, name and the time the bus made it), and
HearEvents and Conn.Consume claim through it, releasing on stop, drain or the
bus ending the loop.

Builds: design 47 H H-A-196
2026-11-20 22:05:14 +01:00
mesh-tools c15460faf5 Merge pull request 'Design 47 H1: a view reader hears an event on its bare subject; an unhandled event is the module's finding (H-A-166, H-A-169)' (#203) from fix/47-H1-sdk-bare-subject into main 2026-11-20 16:55:39 +00:00
jschoubben a5225dfac3 views: wire the unhandled-event count once, and prove it on the bus
The hub's hook, the path the live fault took, had no test that failed
without it, and each Start appended it again, so a second Start counted
every event twice. Start now wires it once per hub and keeps the time of
the first start; the shared-durable integration test holds the reader's
finding of the one event nobody takes, counted once across two starts.

Builds: design 47 H H-A-169
2026-11-20 17:52:48 +01:00
jschoubben e568e6dc63 views: count the events no handler takes as the module's finding
An event a module consumes that no handler takes was said in the journal
alone, so a part that never acted on it (the runner's placing at 17:34) was
seen only by reading the log. The event hub now tells a hook of each such
subject; the process reader, where it holds a findings writer, counts them
in one finding event-unhandled per module (the count since the start and
the last subject) and clears it at its start, since a new build may handle
the event. Without a findings writer the journal line stands alone.

Builds: design 47 H H-A-166, H-A-169
2026-11-20 17:45:27 +01:00
jschoubben 7aae152760 views: hear a seat's event on its bare subject as on its further parts
A seat's contract says an event on its bare subject (the seat mesh-catalog's
registration-changed), while both view readers adopted it on `<subject>.>`
alone, so the event hub acknowledged it unhandled and the part never acted.
The process reader now registers one handler on both forms of the hub (the
two patterns do not overlap, so a message reaches it once), and the stdio
reader subscribes both forms with the one handler.

Builds: design 47 H H-A-166
2026-11-20 17:43:14 +01:00
jschoubben 3667ec96a0 Design 47 H1: proof 2, a two-module change along a dependency edge delivered by its pipeline (H §5 step 7) 2026-11-20 17:06:03 +01:00
jschoubben 237e5fcc91 catalogue: no mention of the seat mesh-delivery-checker is left (CheckerSeat, its baseline word and the exclusion)
Builds: design 47 H §764 (The checker), H-A-57
2026-11-20 04:55:49 +01:00
jschoubben 034b67110b sdk-seat-copies-h1: the copy of mesh-catalog equals mesh-delivery's file at 9a55e6e8 (through C1d)
Builds: design 47 H §745, H-A-42
2026-11-20 04:16:17 +01:00
jschoubben 229e7ea755 sdk-seat-copies-h1: the copy of mesh-delivery equals mesh-delivery's file at c94e8564 (through C3c)
Builds: design 47 H §737, H-A-95
2026-11-20 03:49:40 +01:00
jschoubben 4b7889e0d6 sdk-seat-copies-h1: the copies of git, mesh-delivery and mesh-catalog equal mesh-delivery's files at be741a7a (through C1c)
Builds: design 47 H §737, §745, H-A-42
2026-11-20 03:46:23 +01:00
jschoubben f1b828c6e5 process: Account.NamedEventConsumer binds the durable <node>_<module>_<suffix> beside EventConsumer's
Builds: design 47 H §743, H-A-90
2026-11-20 03:31:31 +01:00
jschoubben 7f3b712a1f motion: a want carries no composed digest, so a have at the wanted fingerprint is applied; the test copy of mesh-wants at C-W
Builds: design 47 H §750, H-A-37, H-A-78
2026-11-20 03:11:33 +01:00
jschoubben 177847e24f sdk-seat-copies-h1: the copy of git equals mesh-delivery's file at d24d6a08; the test build seats take their acts without a bound
Builds: design 47 H §734, §736, H-A-42, H-A-71
2026-11-20 02:53:20 +01:00
jschoubben 67369c99aa sdk-seat-copies-h1: the copy of mesh-delivery equals mesh-delivery's file at C3 (6e027de5)
Builds: design 47 H §737, §752, H-A-42
2026-11-20 02:50:53 +01:00
jschoubben 809fe9aa66 sdk-seat-copies-h1: the copy of git equals mesh-delivery's file at C2 (f8bbde96); the timeout comment names the 30 seconds bound
Builds: design 47 H §734, §736, H-A-42, H-A-65
2026-11-20 02:46:24 +01:00
jschoubben 77a7b187f4 sdk-seat-copies-h1: the copy of mesh-catalog equals mesh-delivery's file at C1 (f006ab50); git and mesh-delivery equal theirs
Builds: design 47 H §745, §752, H-A-42
2026-11-20 02:43:42 +01:00
jschoubben c74f332eba sdk-holder-callers-gone: the gate knows a first-view tool by its name; ViewCallersPrefix and views.CallersEnv go
Builds: design 47 H §752, H-A-66, H-A-69
2026-11-20 02:37:45 +01:00
jschoubben b54a05a687 sdk-holder-callers-gone: no holder keeps a callers list; every seat-list tool and first-view tool answers every caller
go/callers.go goes (stdio.Callers, VerbCallers, ParseCallers, ParseVerbCallers, ConfiguredVerbCallers, HoldVerbCallers, AnyPerson, CallerSeatPrefix; ModuleOf stays in caller.go). seatlist.NewPart(prefix), Tools(r, read, again); ToolsWith, Refusal, RefusalWith and Part.Holders go. views.FirstViewTool loses its callers argument; views.ParseCallers and views.Refusal go. approval.ViewCallersPrefix and views.CallersEnv stay until the question of how the gate tells a first-view tool is answered.

Builds: design 47 H §752, §764, H-A-41, H-A-66
2026-11-20 02:37:10 +01:00
jschoubben 98e3209d70 sdk-seat-copies-h1: the copy of node-engine takes the step placing
Builds: design 47 H §771, H-A-67
2026-11-20 02:36:05 +01:00
jschoubben bc6c7a7710 sdk-callers-refused: seatdef refuses callers and drops the checker's copies
A definition naming callers is refused (a field the format does not know); Callers, DefaultCallers, EffectiveCallers, CallerEntryProblem and the Caller constants go. The test seat list's seats lose their callers. The seat mesh-delivery-checker's bootstrap, test and captured copies go.

The copies of git, mesh-delivery and mesh-catalog (bootstrap, test, captured) have only their callers lines taken out by hand so this commit's tests are green: they are replaced by mesh-delivery's contracts files in sdk-seat-copies-h1's remainder.

Builds: design 47 H §752, §764 (The checker, Every seat contract), H-A-41, H-A-42, H-A-57
2026-11-20 02:34:34 +01:00
jschoubben f60ccc2cfa sdk-seat-copies-h1: the copies of node-engine and operator-channel equal mesh-contracts' files at feat/47-H1-seats
The copies of git, mesh-delivery and mesh-catalog follow with mesh-delivery's contracts commit.

Builds: design 47 H §751, §764, §771, §772, H-A-42
2026-11-20 02:32:16 +01:00
jschoubben bd2db73c3b approval: the gate reads no held-back ask (the operator channel offers every ask)
Builds: design 47 H §764 (The operator channel's limit), H-A-57
2026-11-20 02:31:54 +01:00
jschoubben 59a73e72c0 seatdef-verb-timeout: a verb's definition names a timeout of whole seconds, 1 to 600, others refused
Builds: design 47 H §760, H-A-43
2026-11-20 02:30:41 +01:00
jschoubben be306f4a04 seat list: the captured mesh-catalog and mesh-config definitions lose config-candidates and kept (issue 1982; ADR 0681 §7)
The bootstrapped, test and beside copies are the seats' files as the contracts hold them after the removal of the pull
request path (design 47 G); the views test names a
state the indexer still has.

Builds: ADR 0681 §7 (mesh-sdk's seat list definitions of mesh-config's `kept` and the mesh-catalog seat's `config-candidates` go).

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 22:51:27 +01:00
jschoubben e5d5b2d267 approval: the test table as gofmt writes it, so the sdk's lint step passes and #198's version can be published
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 12:06:12 +01:00
jschoubben 3bb2c4e7f4 catalogue: Resolve keeps the digest of a templated module's archive as TemplatesDigest (A-481, A-481a)
Builds: design 47 G A-481, A-481a
2026-11-17 23:43:08 +01:00
jschoubben 2559c205bb approval: the first views run before any seat list is read (novox 2026-11-16 deadlock)
Cause: startRead looked only in the definitions the runner holds. The seat list is itself read through the seat
mesh-catalog, so before the catalog half is ready the runner holds none, and every first view it waits on (its own
and another runner's, from any account) was refused with "marks not read yet" for ever.

Quick fix: let the one account that stalled through (calledByModule). Thorough fix (chosen): the foundation's bootstrapped
definitions answer for the seats they name wherever the held list lacks them, so a first view runs whoever asks; an act
of the same seat and a seat no definition names stay held.

Test: TestTheFirstViewsRunBeforeAnySeatListIsRead (fails without run.go change).

Builds: design 47 G A-387, A-388, A-433
2026-11-16 18:00:04 +01:00
jschoubben 9b77f82c8b process: the integration test of the seat's callers follows A-434 (an unproven caller of a read definition is admitted)
Builds: design 47 G A-434
2026-11-16 09:40:47 +01:00
jschoubben 88c385a7d2 process: an unread seat definition holds a verb that may need approval for a caller that is no module's account
A-435: the mesh MCP server's account, a person and an unproven caller are held as before while the definition is unread;
a module's account runs it (A-433).

Builds: design 47 G A-433, A-435
2026-11-16 09:40:21 +01:00
jschoubben 1182cadc9d approval: before the catalog half is read, a module account's call runs unmarked, seat verb or tool, and a module's own call in its runtime of its own (design 47 G A-433) 2026-11-16 02:34:22 +01:00
jschoubben 45ee5a2a88 views: a picture whose answer names no make time is judged by its revision alone
The assigner's first view of the indexer's modules stayed behind from 2026-11-14 15:04 UTC: events said the state was
made 2026-10-29, the answer named no make time, and zero is always before it, so it asked 378 times and took in no
event; every want kept its build of that day (the indexer's own new build was never wanted, so mesh-delivery #219
was never taken in).
2026-11-15 23:38:39 +01:00
jschoubben 58c396e3dc process: a seat verb answered by a process admits every caller
The operator, 2026-11-15 23:25 CET: every module, agent and operator may call anything. A marked verb still goes
through the holder's gate.
2026-11-15 23:27:12 +01:00
jschoubben 8eb20000c9 Findings writer: a Write or Clear never holds its caller on the bus past SayWithin, and never holds its lock while waiting on it
The writer's lock was held through the event's resend (up to PublishRetryFor, one minute), so one unacknowledged event
blocked every other Write and Clear behind it.

Builds: design 47 G §383 (the next statement covers an event not acknowledged; the state alone holds the finding)
2026-11-15 01:38:17 +01:00
jschoubben 958b327432 EventHub.StartKept: a program binds its durable on each kept seat event's stream it handles (A-397: the tool runner reads its own twin) 2026-11-14 18:58:20 +01:00
jschoubben 9ec8a88cb0 A call nobody answered names the subjects it was asked on and how long it waited, so its caller's log says where it went unanswered (operator, 2026-11-14: add logging) 2026-11-14 16:17:14 +01:00
jschoubben 7fab6ecde2 Let another module's call of a module tool run unmarked before the catalog half is read
The take-in's call of the forge bridge's tool was refused while the runner's catalog half was
unread: a module tool names no class, and its marks live only in the half the take-in makes.
A module tool whose proven caller is another module's account now runs unmarked then, logged
once per tool with its caller; the console's account (the runner's own), a person, the
controller and an unproven caller still wait and are judged as before.

Builds: design 47 G lead answers A-389

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-14 02:40:37 +01:00