srcOf and the Src and FpSrc stamps go with the source per merged message: a module's value and
its catalog entry carry where it was taken in by its commit alone.
Builds: design 47 H §764 (the indexer); H-A-48
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
I8: the take-in fed by the forge bridge's messages goes with B1's forgemsg: the FORGE
consumer and its handlers, the group end, the pending state with its keys, claims, dry
take-in and hand tools (indexer_pending, indexer_take_in_pending, indexer_drop_pending,
indexer_take_in_merge, kept again for the held take-in in I5), the catch-up by comparison
(kickBehind, TakeInBehind, sayUntaken, takeInHeads, kickCatchUp, CatchUp, merge-untaken),
the source numbering in memory (nextSource), the source verb's follow and Src stamping on
a source answer, the bridge's merged message call, the pull request candidates written
from pull and open messages, the checker's previews and verdict hearing. A repository's
trees at a commit are the internal whole (wholeAt), which ApplyMessage takes in; the dry
indexer stays for a run on request's candidate. The manifest drops the consumes of
forge.**, the checker's verdict and version-delivered, the reads of the checker's
previews and the invokes of open-pulls, bridge_merged_message and mesh-delivery.source.
RoomAgain says room when a take-in is held on a full store (heldFull, I5 holds it).
H-A-82: the fill of delivered versions into the edges goes: HearVersionDelivered, the
delivered.<module> key, the first fill's kinds 2, 4 and 6 from the placement and the
stream running, and outdated's kinds 2, 4 and 6; their entries read not yet delivered.
The tests of the removed paths go (index_test's candidate groups, source_test,
branch_take_in_test, previews_test, candidate_manifests_test, the FORGE handle tests);
the versions tests now hold kinds 2, 4 and 6 as declarations. The integration tests
still drive FORGE and are rewritten in I10.
Builds: design 47 H §743, §744, §764 (the indexer); H-A-48, H-A-82
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
I7: the indexer judges no manifest. The judge pass, its verdicts and unjudged states,
the events manifest-accepted and manifest-refused, the tools verdicts, indexer_verdicts,
indexer_judge and indexer_say_accepted, and the take-in's wait for the verdicts state
go; indexer_bus_state answers from the modules state at its mark. The seed writes
validate required in every policy. oldsections stays (H-A-79).
I6, the half that follows C-R: the gate's pinned, freed, heads-replaced and
pins-stated are no longer heard. A pipeline-started pins the revision it names under
pipeline.<branch>.<source>, its pipeline-ended frees it, each on a durable of its own
stream; kept revisions are judged against the pins the state holds, never a count
(made and pins-stated go). A take-in pins its revision once, for its one pipeline.
indexer_pin_drop (§765) is still to do.
The entry's pull goes (§763). The tests of the old take-in (pending, behind, pushed)
go with it, ahead of I8. The unit suite is green except manifest_test's check that
the watchman's and the grants part's invokes of the indexer's verdicts tools are
served: their removal is the contracts developer's (H-A-79).
Builds: design 47 H §743, §745, §749, §763, §764 (the indexer); H-A-48, H-A-49, H-A-79
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
At its start and on every source revision the indexer compares each repository's taken merge with the head the gate's source names; a repository behind is taken in as its merged message would have been, with the finding merge-untaken raised until it is.
Builds: design 47 G-22 A-463
messageVia, stepsVia, sourceVia and versionsVia go through CallKeptWhileOpen: open while the commit's take-in
is pending, the candidate is held, the branch is watched, the module is in the newest view. A call that ends
with its ask closed is index.ErrAskClosed, not an unanswered call.
Builds: design 47 G §351, §367, §460, §531; A-190
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 node agent calls a restored part's catch-up tool with no arguments, so
indexer_take_in with neither module nor repository takes in the head the
gate's source names for every repository, as the first start's catch-up
does, and keeps the view as the next revision when an entry moved. The
manifest names it beside the catalog graph revisions, form 1 (design 47,
Reverted).
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
main.go reads the view and the module graph's nodes (NodeOf) from the newest trunk revision instead of the modules and
graph states; every write outside a recompute (the judge pass's verdicts, a build's imports, the source stamp, the
modules the trunk dropped, the indexer's own entry a held revision writes) ends in KeepView, so a restart loses nothing;
a revision lists every module the view holds, one whose manifest cannot be read too. At a first start whose objects keep
no trunk revision the indexer takes in the head the gate's source names for every repository by the forge bridge's
merged message and keeps revision 1 (the lead's ruling A). The whole-module tests read the view from the revisions.
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
The catch-up of the heads the source names runs before a source-moved is read, once per branch; the gate's source bucket is no longer read.
internal/index compiles against the new contracts again (SourceByMainValue is gone); tests of earlier subtasks that
the broken compile hid are fixed (the validator's MAJOR.MINOR version in their manifests).
Builds: design 47 G §15, §115, §362
Claude-Session: https://claude.ai/code/session_01JG73KEq4zoDfM9WtYpcQgZ
mesh/check checking edf7b3c0bf3a on b1f262d7d06e | novox/mesh-delivery#129 in group feat/judge-once
Issue 1585, R1: fresh holds the digests a take-in newly wrote (a new module, a changed manifest); a registered or
taken-in module is judged too; a start pass that fails is judged again at the next message heard, no timer.
R2: a module's verdict, problems and warnings are carried while its digest stands and are no part of the compare.
R3: the checker judges the trunk beside the change and fails on every problem the change case adds, an untouched
manifest's included. R4: the judge reads the indexer's kept predefined keys, or the catalogue member's when it
changes them. A log line names a package no held module publishes. Every module pins mesh-sdk 62cddf4.
Claude-Session: https://claude.ai/code/session_01NcbgfJ9Gum47zkFKWRaovL
A registered module whose repository the forge answered 404 (car-hunter, snake, before the forge's account could read
them) was read again only when the indexer started again or the controller wrote a new build record of it, so the grants
part's composed user list stayed short of their subjects and the take-over was refused. A failed read now keeps the
module as unread, read again on the next source revision and the next merged message (no timer of its own), and the new
tool indexer_take_in (the tools bundle mesh-delivery-indexer-tools, through the indexer's takeins state, as the
builder's builder_ask_again) takes one in now and answers what it stored or why not.
The operator, 2026-10-25: only if a module has a new version in commit x is a new value written naming commit x. A merge
wrote every module of the repository at the new commit, fingerprint unchanged (about 186 keys per mesh-catalog merge), and
the controller's take-in took each in again: twice tonight it flooded. Now a merge that gives a module the value it holds
but for the commit writes nothing, and the value keeps the commit its fingerprint first appeared at. The kind: a write
that differs only in where it was computed. The other instance, a source revision re-stamping Src on every module at its
repository's head after any merge anywhere, now stamps Src once. The trunk head stays the gate's source.main.
I1: a merge writes each module its manifests name, at its directory. I2: a pull request's changed directory, a
preview's touched module and a source revision's modules come from the modules bucket (a module a pull request adds is
found in the preview by its manifest). I4: the value of today's form (fp, none) is no longer written, since it was
computed from the build records; mesh-sdk keeps it readable until R4g. One commit: the three share applyMerged and its
tests. Pins mesh-sdk feat/r4e-sdk 9cb4d59.