The runtime's objects delete already purges an object's chunks, and purges
them again when asked again: Purge and purgeObject ask it, and no act of
their own is added.
Builds: design 47 H2 §812; H-A-366 (correction)
The runtime's objects act purge removes a deleted object's chunks:
busstate.Objects.Purge and TypeScript purgeObject, so the registry's
program opens no connection of its own for it.
Builds: design 47 H2 §812; H-A-366
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
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
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
Main reverted G's half merge (#189, reverted by 8afc3eb in #191) and then took the reader release's S1-S5, which
copied G's code. G's form holds where they differ (process_bus.go, seats.go, state.go and their tests); the
release's kept names stay so main's callers keep building: SeatViewPrefix beside SeatReadPrefix, and QueueInvoke's
empty strings on not-ok (with its test TestQueueInvoke). The tree differs from G's head f30dee54 only by those and
the release's two tests (objects cap option, seat states and stream kinds).
Builds: design 47 G (the milestone's sdk part, as at f30dee54)
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
mesh/check being carried to 8afc3ebdf371 on 93e0a7b50d90 | from 27da949140c6
The group feat/47-G-join11 merged 3 of 17 members on 2026-11-13; G's mesh-sdk alone on main breaks 25 trunk
modules. The operator chose to put the order right: a reader release first, then G whole. This restores main
to 8fbba810 exactly.
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
views.FindingsAskedSubject, views.FindingsAskedEvent and FINDINGS_ASKED_EVENT are gone; the stdio writer subscribes the seat event
the given subject names.
Builds: design 47 G A-17, A-13, §461
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
mesh/capabilities answered {words}, an unknown call read as none, the node from MESH_NODE; the finding writer's bundle
transport requires finding-calls; a runtime refusal keeps its code.
Builds: design 47 G §576, §622, §488
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
The TypeScript writer holds no FindingBody of its own and works on the wire form (checkFinding, views reader findings laid out by
viewFindingBody); both writers name findings-asked as the seat mesh-conditions' event. The view-waits test waits on the raise,
not on a count of yields.
Builds: design 47 G §686, §640, §687, A-3, A-13, A-17
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
As the Go reader's DefaultCallers: read takes invokers, persons and sessions,
act the operator and sessions, warrant the operator; no class reads as read.
Builds: design 47 G §391, §456, §196
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
The forge module reads the catalog half from TypeScript (W12.7) and must wait and say so as the Go readers do (design 47,
A part with no view says it waits): views' Follower gains waitFor and holds, raiseWaits and clearWaits, the status fields
view, waits-for, waiting-since and waits-raised, WAITS_BOUND_MS (five minutes) and viewWaitsFinding, by the Go side's
rules (go/views follow.go, findings.go). Rows in the first-view test of test/sdk.test.ts.
Builds: design 47 G §196, §200
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
A TypeScript holder, the forge module first, must judge a verb's caller by the runtime's proof, say its seat's events
and answer its first views' read and ask-again tools (design 47, The forge module's tools come from mesh-sdk's
TypeScript side; G-F1 option A). The stdio session now hands each tool the caller and call id the runtime wrote in
_meta (callers.callerOf, the Go rules for Callers and VerbCallers), binds the runtime for seatPublish over
mesh/seat.publish and config, and views gains the Follower and its two status tools over the state first-views.
@novox/mesh-sdk 0.1.19.
Builds: design 47 G §196, §200, §336
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
mesh/check pass 18e541c1badc on 7187beaf042b | carried from 2cbe21937153 | a compiled program is not refused: mesh-build-node reported no programs | u…
process.Conn.Keeps, stdio.Keeps and the TypeScript keeps publish the announcement on
the module's own subject; afterwards a state the code did not announce is refused at
the call, so announcement and code cannot drift. The package goes to 0.1.15 for keeps.
mesh/merge-gate head 1bc5719b8b85 merged into trunk a2d2af00dc5d beside mesh-host#116@bff1ac14 — pass, with a note: builds mesh-sdk (+40 dependent(s)) → ac…
mesh/repo-check head 1bc5719b8b85 merged into trunk a2d2af00dc5d beside mesh-host#116@bff1ac14 — pass: its merge-check.sh passed, each part in its toolchai…
mesh/merge-gate head d49d3a14d82a merged into trunk c1ae30b0a2c8 — pass: builds mesh-sdk, sent nowhere; no bus step; every machine composes with the change…
mesh/repo-check head d49d3a14d82a merged into trunk c1ae30b0a2c8 — pass: its merge-check.sh passed, each part in its toolchain: merge-check.sh (typescript)…
No provider on runProvisioner passed announce, and the loop had no standing per
consumer, so the object store, the SQL and document stores, the mail server and
the others failed their consumers in their journals alone. The loop now keeps the
Go copy's standing (failing after five minutes, again every fifteen, recovered on
success, when unasked or retired, and on the first success since start) and says
it through the SDK's emit unless given another announce.
A handler takes an event by returning and asks for it again by throwing;
any mesh/event answer but an error is taken. Written down in the README
where event handling is, with what follows for a handler: throw only when
the work was not done, and be safe to run twice.
Both SDKs added the handler before asking the runtime and kept it when the
runtime refused, so a module that asks again until it is taken (every Go
consumer does) ran its handler once per attempt for every event. A refused
subscribe now removes its handler; one the Go runtime did not answer in
time is kept, since the runtime may have bound it.
Five passes are 25 s, shorter than a real hiccup. And a provider's remove can
destroy — a ledger file, a public DNS record — so an adapter without a retire
of its own is mark-only: the consumer keeps its access, is marked and said, and
remove is reached only by a person's delete (hq ADR 0230).
The hourly release let a mistaken bulk unassignment drain one consumer an hour
with nobody deciding. Per novox/hq ADR 0230 a consumer is retired (disabled,
marked, data kept) only after the same unasked set in five consecutive passes;
more than 3, or more than half of those held, waits for approve or reject.
Deletion is a separate act of a person, of a retired consumer only.
Issue 241 withdrew seven consumers in one pass; 0.1.10 refuses a file it
cannot read, and a file read whole that names nobody still withdrew
everybody. A pass that would withdraw more than withdrawAtOnce consumers
(1), or more than withdrawFraction (0.5) of those held, now withdraws
nothing and says so; while the mesh goes on not asking, one is let go every
releaseEveryMs (an hour). A consumer asked for again is kept.
An unreadable, unparsable or unrecognised contributions file was read as "no consumer asks", and the
reconcile pass then removed every consumer it had applied. On 2026-10-04 one such read made the
postgres provider drop every database on the control node. Such a pass now applies and removes
nothing and says why, once; only a file read with a given list withdraws a consumer, and every
removal is logged before it is made.
The harness says 'secret not readable yet' when it cannot open a consumer's
sealed credential, which is true and ordinary on the first pass. It said it
4330 times in three hours on the control machine while two consumers were
never provisioned, because a permission the host will never grant reads
exactly like a file that has not arrived.
After about a minute it says so plainly instead, once and then rarely, and
names what to look at: who owns the file and who this process runs as. The
ownership itself is the controller's to fix.
0.1.9.
state(name) / stdio.State(name): get, put, delete, keys and watch over the
runtime's mesh/state verbs. A watch's handler is registered under the id it
is asked with, so its current values — which arrive before the watch is
answered — and its changes reach it and no other watch.
Provision gains `derived`; the harness keeps it per applied login so remove()
is told it too, and includes it in the change hash — a provider that renames
what it derives has given the consumer a different resource.
Version 0.1.7. Rebased onto main after the bundles refactor; the record this
cites was written as 0188 and renumbered to 0201, because ADR 0188 on main is
now the bundles record.
The stdio channel is the bundle's bus: subscribe asks the runtime (mesh/subscribe), which binds the
module's durable consumer and hands each event over as mesh/event — answered once every matching
handler has taken it, as an error when one failed so the runtime does not acknowledge it; request
asks the runtime (mesh/ask). The session starts as the stdio module loads when the runtime launched
the process, because a handler module subscribes at import, before the launcher says what it
serves; initialize and tools/list are answered once it has. topicMatches as the runtime's. 0.1.6.
A launched bundle holds no bus connection. The stdio loop binds a broker whose publish asks the
runtime (mesh/publish) and resolves when the runtime has published; asking, answering and
subscribing are refused by name. Tool calls run beside the loop, so a tool awaiting the runtime's
answer to its emit does not deadlock the line that carries it. 0.1.5.
serveRegisteredOverStdio lists the registrations of the module named by MESH_SERVED_MODULE by
their own names and every other registration as <seat>.<verb>. Before, the first registration
was taken for the module's own, so a bundle registering its seat first listed the seat's verbs as
its own tools. Contributors are given the process's environment, the bundle's own. 0.1.4.
A node's runtime serves many modules in one process; a module's composed paths and words must
reach its own tools and no other's. The runtime knows which bundle made which registration, so
it says which environment each contributor gets. collectTools(env) is unchanged. 0.1.3.
The node's runtime launches a tools bundle as a process and speaks MCP over stdio to it, so a tool
may be written in any language and the transport stays in the runtime (ADR 0039's refusal, kept).
Each SDK is the loop and the tool type and nothing else: `@novox/mesh-sdk/stdio` (serveStdio, and
serveRegisteredOverStdio for a bundle that already registers the in-process way), python/, go/,
rust/ and c/ beside it, each with an example answering `greet` and the seat verb `node-lamp.on`.
Skeletons, by the operator's direction: the bare minimum one bundle per language needs to be built
and answer, proven here by each example answering tools/list and tools/call on stdio.
A lost consumer whose re-apply fails is retried at the next check with
its count unchanged, instead of waiting out a backoff meant for adapters
whose create and holds disagree. A check that fails for one consumer no
longer stops checking the consumers after it; only a timeout does.
A check that hangs no longer stalls every consumer: it times out after
30s and counts as could-not-ask, and the rest of that pass is not asked.
A consumer still not held after being applied again is checked at
doubling intervals up to an hour, and said loudly, so an adapter whose
create and holds disagree costs one re-apply an hour, not one a minute.
The consumer's password is scrubbed from every error the harness logs.
An optional holds() on the adapter is asked for every applied consumer
every minute; false applies it again. A backend that forgets what it was
given while the provisioner runs (hq issue 120) is healed within a
minute instead of failing its consumers in silence. Unable to ask is not
treated as loss. Adapters without holds() behave as before.
One name per thing, per the HQ glossary: the module/container/image/binary/repo
becomes mesh-controller, the seat the-controller, and the store+broker pair the
foundation (embedded base bundles, default template and example lock renamed with
their go:embed directives). No behaviour change — a pure vocabulary rename.
Claude-Session: https://claude.ai/code/session_01D6qtiYU3P9jk3pnAXyAFyx