Design 47 G start: catalog fixes placed by hand and the Go build per module (A-447) #691

Merged
mesh-tools merged 26 commits from fix/47-G-start-merge into main 2026-11-16 14:15:15 +00:00
26 Commits
Author SHA1 Message Date
jschoubben f31d6c645d Merge commit 'cdf16870' into fix/47-G-start-merge 2026-11-16 15:11:01 +01:00
jschoubben bbc9eb7c93 Merge remote-tracking branch 'origin/main' into fix/47-G-start-merge 2026-11-16 15:11:01 +01:00
jschoubben cdf168706b Go steps test a module's own directories alone, not what its go.mod replaces from outside (A-447c)
Builds: design 47 G (G-22 lead answers 2) A-447c
2026-11-16 11:42:56 +01:00
jschoubben ae78fb039f Go fetch with no dir fetches the tree whole again, for the checker until G removes it (A-447e)
Builds: design 47 G (G-22 lead answers 2) A-447e
2026-11-16 11:40:18 +01:00
jschoubben 928de07c43 Go build module: a step for a module compares its own module.json's generated files alone (A-447c)
Builds: design 47 G (G-22 lead answers 2) A-447c
2026-11-16 11:37:23 +01:00
jschoubben 666be37916 Go fetch is refused with no dir, naming its asker; the checking program's fetch names its directory (A-447b)
Builds: design 47 G (G-22 lead answers 2) A-447b
2026-11-16 11:34:17 +01:00
jschoubben ccab80d726 Go fetch with dir: a directory with no go.mod chooses every go.mod under it; none chosen fails naming it (A-447a)
Builds: design 47 G (G-22 lead answers 2) A-447a
2026-11-16 11:16:16 +01:00
jschoubben 1168376930 Go build module: a fetch asked with dir chooses that module and its local replaces alone (A-447)
Builds: design 47 G (G-22 lead answers 2) A-447
2026-11-16 11:09:26 +01:00
jschoubben d2c1a24a24 nats: the bus keeper no longer reads the seat mesh-catalog's states before the first views (reverts ed536102)
A-379 replaced A-376 for the start, and the seat mesh-catalog offers neither the states modules and manifests as views
nor verbs of those names, so the validator refused the manifest's two consumes and two invokes.

Builds: design 47 G A-379, A-445 (G-22-lead-answers-2.md)
2026-11-16 09:51:48 +01:00
jschoubben 88ff34e855 distribution: a step given up at its bound is said given up, and its call's end is said later
Builds: design 47 G A-443, A-443a
2026-11-16 09:46:12 +01:00
jschoubben 76195e421f mesh-delivery-gate: the start pass runs once, asking again only while the registrations cannot be read
Builds: design 47 G A-442, A-401
2026-11-16 09:41:45 +01:00
jschoubben 59704456c3 messenger: offer every ask when it is raised; hold none back for room
The operator channel kept at most three asks of one asker offered and held its oldest back behind newer ones, so asks
waited unseen (the pipeline's proposal behind two mailu asks). It keeps no count any more and holds nothing back: every
ask is offered when it is raised, and lapses at its expiry or ends at its answer. An ask held back before the change is
offered again at the router's next tick; ask-again offers one so and holds nothing back in its place.

Builds: design 47 G A-424
2026-11-15 15:52:48 +01:00
jschoubben 4a99f7044e distribution: a publish with a version and no parts tags the digest the store holds
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
2026-11-15 02:39:28 +01:00
jschoubben 7d17441a9a gofmt 2026-11-15 01:38:29 +01:00
jschoubben c006538064 Distribution: the incoming-full finding is said off the publish's path
Builds: design 47 G (publish path of oci-artifact-registry; no design line covers it)
2026-11-15 01:38:27 +01:00
jschoubben bfc877209f Distribution: each bus step of a publish bounded at 10 s and logged by name
Builds: design 47 G (publish path of oci-artifact-registry; no design line covers the bounds or log lines)
2026-11-15 01:30:05 +01:00
jschoubben 99f3e23492 Distribution: each wait of a publish logged with its time; the window's learning no longer holds the answer
Builds: design 47 G (publish path of oci-artifact-registry; no design line covers the log lines)
2026-11-15 01:20:04 +01:00
jschoubben e8204a9c27 Gate: the start pass asks again until no repository is left out (A-401)
The forge's git verbs may serve after the gate does; one pass with every head read refused wrote nothing and was never retried.

Builds: design 47 G A-401
2026-11-14 22:34:54 +01:00
jschoubben 89397466aa Gate: source-moved for started heads names no group; tests for the registration-changed branch and the start pass (A-401, A-404)
Builds: design 47 G A-401, A-404
2026-11-14 22:16:30 +01:00
jschoubben 28f616e379 Gate: a held source key gets the registered repositories it lacks, heads unchanged (A-403)
Builds: design 47 G A-403, A-401, A-34
2026-11-14 22:09:06 +01:00
jschoubben 8ea8ef9768 Gate: a watched branch starts from the forge, as a clone does; tool gate_source_start (A-401)
At its start and on each registration-changed of the seat mesh-catalog, the gate writes source.<branch> for every watched
branch the registrations name that no source key holds, each registered repository at its head from the seat git's verb
branch, group empty, then says source-moved; a repository whose head cannot be read is left out and said; a held key is
never rewritten. gate_source_start (branch, why) does it by hand through the hand-act log. The manifest consumes
registration-changed and invokes seat:git.branch and seat:mesh-catalog.registrations.

Builds: design 47 G A-401, §430, §15, A-34
2026-11-14 22:08:22 +01:00
jschoubben fb24d384a2 Build seat: says each toolchain-reported it said, not only the ones it could not 2026-11-14 19:00:01 +01:00
jschoubben 5236c2aa72 gitea: the seat git's branch with no branch named answers the repository's default branch (A-399)
The seed and a registration with no branches read the default branch through the seat git (design 47 G-1 §104) by
calling branch with an empty branch; gitea refused it as no branch for a path, and the seed stopped on novox 2026-11-14.
2026-11-14 14:57:27 +01:00
jschoubben 5df0fb64f5 mesh-build-go: a finding writer that cannot start is said, and the module serves on (A-394)
Live on shanks the bundle exited 39 times on 'an ordered consumer of mesh-conditions.findings-asked: context deadline exceeded', so the seat node-build-go never answered steps and the seed waited.

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-14 11:52:23 +01:00
jschoubben ed536102ac A-376: the bus keeper makes the catalog half's own holder before the first views
Before the catalog half's first views are in (a read refused, or none-yet), the keeper reads the manifest on the trunk of the
holder of the seat mesh-catalog from the seat's states modules and manifests (the holder from the seat that places modules)
and makes its kept data as any module's, each entry said as ever; every other module waits as A-367 says.

Builds: design 47 G A-376 (G-22-lead-answers-2.md)

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-13 23:11:57 +01:00
jschoubben a98b6ba77e S6: the bus keeper treats a none-yet catalog half as not yet viewed
mesh/delivery ready: it is merged on the Approve
mesh/delivery-group group fix/47-G-start-2 checking: 2 of 2 member(s) ready
PlanCatalog plans nothing, deletes no plans key and orphans no entry while the
catalog half lists no module before any was seen; the first revision with
entries plans as before.

Builds: design 47 G A-367 (G-22-lead-answers-2.md), W22/Q4 of start-waits.md

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-13 20:48:04 +01:00