Held compared the version tag's manifest digest with a bundle's or an
archive's blob digest, so every second publish of such a version was
refused even at the same bytes. It now compares like with like: a bundle's
or an archive's stored layer, an image's manifest. The same digest is
already held (ended, published said, nothing stored), on a repeated ask too;
another digest is refused naming the held one. The fake store keeps a
manifest digest beside the layer's, and a row runs the module's OCI code
over a store in memory.
Builds: design 47 H2 H-A-474, design 47 H H-A-393
The merge of main's H-A-387/H-A-393 fix gave Store.Held a reference and the
digest another publish holds; a repeat answers ended when the reference is set.
Builds: design 47 H2 §775; H-A-381
While the reserved digest is still empty, a second publish of the same
version with another digest passed the reservation check and the store's
Put moved the version's tag. Publish now refuses it in its held branch,
before the reservation is asked, naming both digests, with the publish's
finding and refused said as any refusal; the same digest is taken as before.
The store's Held answers the digest it holds at the version when it differs.
Builds: design 47 H H-A-393, design 47 G A-490j
A publish carrying an ask id is kept in the state taken for ask-taken-bound
(10m): a repeat while the act runs joins it, a repeat after it ended answers
the act's state read afresh (the transfer's lacking parts, a held version
ended, a transfer gone refused), never publishing twice.
Builds: design 47 H2 §775, §820; H-A-359, H-A-381
The builder's publish again of mesh-delivery-pipeline 0.1.6 sends the version and the digest with no parts, as A-73
has it; the build stored the bundle untagged, and the holder answered 'publish: no parts' unless the tag existed
already. The holder now checks the digest against the store under the name (the artifact store, A-50), asks the
builder's reservation, tags it with the version and says published; a digest it holds none of is refused as not
stored, naming the digest.
Builds: design 47 G A-417, A-73, A-50, §588
The config `callers`, the tool registry_callers and its invokes, the caller-list code and the holder's per-verb
Allowed checks go. The windows' driver counts as build modules the modules claiming a `node-build-*` seat in the
catalog half, followed by the seat mesh-catalog's branch follower on the program's views reader (one shared events
hub); it wakes again when a build module's entry moves. The retention check of remove finds the artifact sweep's
holder through the seat mesh-module-assigner's holders (seatlist.HolderOf), not a name.
Builds: design 47 G §610, §7, §428, A-156, A-82
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 container image and OCI artifact seats' holder asks the seat mesh-builds' verb reserved in place of builder_versions; the manifest
invokes it and takes no builder package.
Builds: design 47 G §381, §478, §504
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
catalogue's SeatReference, InRegistrySeat, Routed, Recorded, InArtifactStore and ArtifactSweepSeat leave mesh-sdk;
each package gives the two registry seats' names from their contracts (or the build step's copies of them).
Claude-Session: https://claude.ai/code/session_01NcbgfFSnn24M3GcqttoU7i
Saying removed for a version the store still holds told the builder something false. remove now answers
kept with the reason shared-digest and the other tags naming its digest, removed false, and says the
seat's event kept with them (mesh-contracts' shared registry part); the builder raises no retry.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
registry_incoming, npm_incoming and go_incoming dropped a transfer by hand with only a log line. Each
drop now records its hand act through the seat mesh-hand-acts before anything goes, as nats_requests
does; an act not recorded leaves the transfer open. distribution's manifest invokes the seat.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
Both holders carried the same idle bound and the same read of incoming's cap; the second instance of the
kind moved into mesh-sdk's Receiver (Arm, Expired, Bound, with BusArm, BusCap and FollowIdle). Each
holder now hands its idle state and its objects to those, and keeps its rows.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
The holder counted against 20 GiB while the bus caps incoming at 8 GiB, so the bus refused writes before
the holder said refused naming the cap, and the 80 % finding came too late (running review F2). The cap is
read from incoming's binding (the object store's fill); IncomingCap, 8 GiB, is said only where none is read,
and the manifest's row holds it equal to max-mib. An abandoned transfer was dropped only on the next publish,
and the transfers state's own lifetime dropped its key while its parts stayed: transfers keeps for good, and
each open transfer has a key in the new state idle with its own time to live, whose end drops it (re-armed
for the rest when it took a part since). registry_incoming reports the bound cap.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
The reach came from MESH_REGISTRY_REACH, which nothing sets, so every published event
answered /<name>@<digest>. The store answers the path; the holder puts the seat's name
before it, as catalogue.InRegistrySeat reads it back (ADR 0605 §9; lead's ruling on A3n).
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
A second tools process following mirrors would copy again and say published twice. The ask is claimed by
compare-and-set before it is copied, and its outcome carries the ask's id as the event id, so a redelivery is one
message. The cap's publish-refused finding stands on incoming, not on one file, and clears with the next publish
taken.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
The holder's own transfer logic goes: parts.Receiver takes the chunks into incoming, checks the hashes, drops an idle
transfer and tells the fill; distribution gives it the builder_versions check, its store and its seats' events, so
the forge's two package seats bind the same receiver instead of a copy.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig
The holder answers publish and publish-part into incoming (parts of at most 1 MiB, resumed from what is held,
hashed, only the missing layers sent, 20 GiB with its 80 % finding and a refusal at the cap, a transfer idle 24
hours dropped), asks builder_versions before it stores, and says published or refused (design 51 *Publishing*,
ADR 0605 §4, ADR 0611). versions, remove (a tagged version; untagged only for the sweep's retention, which a
hold keeps), hold and release (ADR 0617 §2: the sweep alone; the operator's session by hand; never
distribution's own account), tags and mirror. The caller list is a trusted config value checked against the
caller the tool runner proved. The store's port is the program's: reads open, writes refused but while the
direct-write window is open (never over a tag published through the seat) or with the bootstrap's one-off
credential. The old-name and direct-write windows read every node, away ones included, close once, and name a
node holding them seven days.
Claude-Session: https://claude.ai/code/session_01JG73KJNrH5EgFXMKKYSrig