Commit Graph
2634 Commits
Author SHA1 Message Date
jschoubben abdd70d918 I2b: one catalog revision per take-in, <branch>.<source>, carrying its movers
TakeInSource takes a gate's source-moved in: the trees of each repository it moves at the
named commit through the seat git, the fingerprints of the modules registered for it, the
movers against the branch's newest catalog revision, and one revision <escaped
branch>.<source> for every branch, the trunk included, numbered by the source and written
once by read before write. The revision carries its movers, and a redelivery whose revision
is kept says take-in.taken again from them, recomputing nothing (H-A-92). It is pinned for
the take-in's pipeline as pipeline.<branch>.<source> in the same write when it has a mover,
none for a take-in with no mover. take-in.taken carries branch, source, revision, call and
the movers' entries alone, under a duplicate id of its branch and source (H-A-31). A
source-moved with no call is refused as the old form (§765).

Only a take-in keeps a catalog revision (H-A-49): the recompute, the build ends' imports and
the start's reconcile keep none, and KeepView, keepEntries, keepTrunkRevision and
PinPipelines go. Another branch than the first is taken in on its newest revision's view.

Builds: design 47 H §743, §749 (the pin in the take-in's write), §765 (the old form); H-A-31, H-A-32, H-A-49, H-A-92

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:56:27 +01:00
jschoubben d00ee48363 I2a: the movers of a take-in are the moved fingerprints and their importers
A take-in's movers are the modules registered for a repository the source revision moved
whose fingerprint differs from their entry in the branch's newest catalog revision, and
every module, in any repository, importing a package one of them publishes. A module of
another repository is no mover by its fingerprint alone; one with no build is none.

Builds: design 47 H §743 (movers); A-476a

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:56:27 +01:00
jschoubben 1e826ff008 I8, H-A-82: the earlier take-in goes; no delivered version is kept
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
2026-11-20 04:56:27 +01:00
jschoubben ea28f39319 I7, I6 (the pins' events): no validation at the take-in; a pipeline's events pin its revision
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
2026-11-20 04:56:27 +01:00
jschoubben 930d09cadc H-A-66: the indexer's revision verb, first views and seat-list part keep no callers
Every account may call every verb: the verb revision answers every caller, so its
callers list and the node-engine account's mapping go; the first-view tools take no
callers list and the tools bundle's env carries none; the seat-list part is made
without one.

Red only from the contract removals C1 and C3 (I7's manifest events, the gate's pin
events, the entry's pull), followed in the next commits.

Builds: design 47 H §752, §764 (every seat contract: callers); H-A-41, H-A-66

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:56:27 +01:00
jschoubben cbf1dbb15d I9: register, watch and the seed read no branch's protection
The forge's branch protection is no longer a condition of the mesh: register and watch
refuse only a reserved branch name, and the seed leaves out no module for its default
branch's protection and reads none.

The module compiles again only once C0 to C3's removals are all followed (H-A-66's
callers, I7's manifest events, the gate's pin events, the entry's pull); its unit
suite runs green from that commit on.

Builds: design 47 H §763 (G §5), §764 (the branch protection)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:56:27 +01:00
jschoubben f9e2dc38d8 W1: the watchman's whole test: a pipeline waiting behind another raises nothing
The indexer holds a source and takes it in, the runner writes its pipeline
waiting its turn (wait-held, step pipeline, on pipelines-at-once) ten minutes
back with no step asked, and the pipeline ahead of it stands: no take-in-waits,
no pipeline-stands and no head-untaken name the waiting one.

Builds: design 47 H §741, §748

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:51:07 +01:00
jschoubben 94d3f02f76 B-src: the builder's floor wakes on the indexer's take-in.taken
The seat mesh-catalog defines no manifest-accepted any more: the builder
consumes the indexer's take-in.taken in its place (module.json, names.go's
patterns) and wakes its floor on every one, looking at every module again as
before (versions.go, the wake take-in.taken). With the in-progress commit
before it the builder reads no source stamp and its untagged suite is green.

Builds: design 47 H §764 (the indexer: its events manifest-refused and manifest-accepted, with their reader); lead answers H-A-48, H-A-49, H-A-66, H-A-116 (hq 508faf0c)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:48:48 +01:00
jschoubben 11a22fdff7 X1: the checker goes: its module, its seat definition and its readers
The module mesh-delivery-checker goes whole with its manifest, and its seat's
definition leaves the seat list. The status view reads no tries: build_usage
answers no tries running and counts no step as a try. The grants part grants
main's accept subject to the builder alone. The tests name no checker as an
asker or a seat. The seat's generated packages, the indexer's readers and the
assigner's verdict constants go in the join's cleanup, once those branches
stand on this one (the lead, 2026-11-20).

Builds: design 47 H §764 (the checker (H1): mesh-delivery-checker whole); lead answer H-A-57 (hq f599b761)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:47:16 +01:00
jschoubben 90f53608bc B-src (in progress): the builder reads no source stamp
A module value carries no source stamp (C1d): ModuleOf takes none from the
entry's inputs, a build's claim is numbered per fingerprint naming no source
revision and no merge, and the log's started entry names none. The seat list's
parts take no callers and resolve no caller's holder (mesh-sdk's NewPart without
callers, SDK-C). internal/... is green; cmd/mesh-delivery-builder still hears
manifest-accepted (names.go:21, versions.go:192, module.json:57), which the
seat mesh-catalog no longer defines: its floor's wake is H-A-49's, asked of the
lead.

Builds: design 47 H §764 (the indexer: the field pull, the source numbering); lead answers H-A-48, H-A-66 (hq f599b761)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:43:10 +01:00
jschoubben 2bb08b9306 G1: the grants part's tests read no verdicts and no callers
The root package's tests and the part's whole test no longer write or read the
indexer's verdicts: a value whose verdict field says refused is composed as
ever (a module in a revision is accepted by being there), and a held module is
one whose manifest the part cannot read. The verdicts' first view row reads the
four first views; the no-verdict window and the want's newest build rows go; the
compare rows carry the denies every account now has, being granted every verb of
every seat; the seat list's again tool's caller row goes with the callers; the
captured nats manifest consumes catalog-revision, as the catalog's does, not the
retired manifest-accepted.

Builds: design 47 H §745, §752, §764 (every seat contract: the callers lists); lead answers H-A-66, H-A-79, H-A-113 (hq f599b761)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:39:46 +01:00
jschoubben ffb5c48dd2 SDK-C: a composer case carries no verdicts
The composer judges no manifest and the assigner's composition step sets
no verdicts, so Case.Verdicts goes with its words; a case carrying
verdicts was read before and is refused by a strict reader now.

Builds: design 47 H §745, §764 (the indexer: the validation at the take-in); lead answer H-A-112

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:28:59 +01:00
jschoubben 6b43adafca G1 (in progress): the grants part reads no callers and no verdicts
Every account may call every verb: the grants part reads no definition's
callers (callerOf, the verb and accept callers, admits) and grants every
module's account every verb of every seat (mesh.seat.*.tool.> and its
caller-named form), a sent store to every module invoking a verb of its
claimant's seat, and every queue's named ask to every asker. grants_callers
goes; grants_why_refused answers the bus's own verdict. The indexer keeps
no verdicts: their first view, events, invoke and the engine's read go, and
a want names its build alone (no newest). internal/... and cmd/... tests are
green; the root package's grants_test.go still reads the verdicts (rows at
grants_test.go:303-307, 741, 1184, 3700-3870, 4560-4615) and does not compile.

Builds: design 47 H §745, §750, §752, §764 (every seat contract: the callers lists); lead answers H-A-37, H-A-66, H-A-79, H-A-113 (hq f599b761)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:28:04 +01:00
jschoubben f8eea0000a V1: the view reads the H1 want, source, source-moved and held take-ins
A want counts as applied by its fingerprint and a motion event is keyed by
it; the want's newest and rule, and the state held back read from them, go
with the assigner's holds. A branch's source is {v, source, call, heads}
with no group, its source-moved is read by its source, and the delivered
event names no group. indexer_pending's held take-ins are said in the
tiers' pending words with their branch, source, why and since, naming no
module. The view reads no source stamp on a module. Its seat list, its
first-view tool answers and its manifest name no callers; its source invoke
reads call, not group.

Builds: design 47 H §737, §744, §750, §752, §764 (the gate's groups; the indexer's pending and groups; the assigner's holds; the callers lists); lead answers H-A-26, H-A-37, H-A-48, H-A-66, H-A-78, H-A-102, H-A-109 (hq 3433db0e)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:20:33 +01:00
jschoubben 9a55e6e8c0 C1d: the seat mesh-catalog's module and inputs carry no source stamps
The indexer stamps no source revision on a module (the take-in's commit
names it), so the seat mesh-catalog's state modules value and EntryInputs
lose src and fp-src, and the fixture rows their stamps. A module value or an
entry's inputs carrying either was read before and is refused now.

Builds: design 47 H §764 (the indexer); lead answer H-A-48

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:15:20 +01:00
jschoubben 78e169a8c6 C-R5: the runner's pipeline keeps the movers superseded out of it
The seat mesh-pipeline-runner's Pipeline gains superseded, the modules a
newer waiting take-in of the same branch took out of the pipeline, written
with the movers by compare-and-set, so the runner writes a mover's
build-step why from it at the pipeline's placing. A fixture row of a waiting
pipeline naming one failed to read back before the field was defined.

Builds: design 47 H §748; lead answer H-A-107 (hq fabb927f)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:11:18 +01:00
jschoubben f6163349af W1: the watchman reads the H1 take-in, the mover's entry and the want
The watchman follows the one model in one pass. Its take-in waits are the seat's take-in.pending
and take-in.taken, keyed <branch>.<source>, timed from the message's own time by the setting
take-in-bound (60 minutes by default), and seeded from indexer_pending's held take-ins; the group
bound goes. A taken-in mover's bus state is read from its catalog entry by fingerprint through
the seat mesh-catalog's verb entry, bus-state-overdue timed from the take-in. The manifest
verdicts and their findings, the head timing of source-moved, acting-part-late and
delivery-waits go, with their readers, consumes and the indexer_verdicts invoke. A want awaiting
its hand placement is not timed, and the want readers follow the H1 want. A record-opened is
heard on record-opened.<call>.

Builds: design 47 H §741, §744, §764 (the indexer: manifest-refused, manifest-accepted, the
watchman's findings; the assigner: acting-part-late, delivery-waits), §772; lead answers H-A-25,
H-A-37, H-A-49, H-A-70, H-A-78 (2), H-A-79, H-A-91 (2), H-A-94, H-A-95 (1), H-A-97, H-A-102

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:05:11 +01:00
jschoubben 5bf6fd9b01 graph: the refused cycle is shared code, read by the indexer and the validate step
RefusedCycle, with its running-version break and its path search, moves from the indexer's
internal/index into mesh-delivery-sdk/graph, so the assigner's validate step reads the same rule
the indexer's take-in does; the indexer calls the shared function. The rule's table test moves
with it; the indexer keeps its test of the trunk catalogues against the shared rule.

Builds: design 47 H §745; lead answers H-A-34, H-A-104 (2)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 04:00:13 +01:00
jschoubben 89e0926feb C-R4: the runner's rerun, accept and ask-again name the mover's module
A pipeline of several movers is never ambiguous: the seat mesh-pipeline-runner's verbs rerun,
accept and ask-again gain the optional input module, the mover's; a call naming none on a
pipeline of one mover is taken as that mover. Fixture rows of each verb's input name a module
and failed to read back before the field was defined.

Builds: design 47 H §746; lead answer H-A-101 (1)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:52:48 +01:00
jschoubben c94e8564e6 C-R3: the runner's queue keeps the pipelines running
The seat mesh-pipeline-runner's state queue gains the key form running.<pipeline>, value
{pipeline}, written when a pipeline is placed and deleted at its end, its cancel or its approval
wait, so pipelines-at-once is judged against the running keys alone. A fixture row reads the new
form; the generator refused it before the form was defined.

Builds: design 47 H §748; lead answer H-A-96

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:49:06 +01:00
jschoubben 969caea37f C3c: the seat mesh-delivery's owed acts keep no merge's ask or group
The Owed definition's fields ask and group were kept only for a merge, and the owed act merge
went in C3b; nothing reads them, so they leave the definition and the generated contracts. A row
of the client's test reads the definition and fails while either field is in it.

Builds: design 47 H §764 (the forge bridge: the merge act of a switched key); lead answer H-A-95 (2)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:48:23 +01:00
jschoubben be741a7a8f C1c: a catalog revision carries its movers
The seat mesh-catalog's CatalogRevision gains movers, the take-in's entries
({module, commit, fingerprint, repository}), so a redelivered source-moved whose revision is
already kept says take-in.taken again from them, recomputing nothing. The branch revision's
fixture names one mover; the generated fixture test fails without the field.

Builds: design 47 H §743; lead answer H-A-92 (hq 8c0b04e3)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:40:18 +01:00
jschoubben f79c0a7b78 C3b: the seat mesh-delivery's owed acts name no merge
The bridge performs no merge since the merge call (design 47 H §764), so the owed act merge
(OwedActMerge) leaves the seat's definition, its fixture and the generated contracts; the
bridge's tests no longer read a merge fixture to refuse. A row of the client's test reads the
definition's act enum and fails while merge is in it.

Builds: design 47 H §764 (the forge bridge: the merge act of a switched key); lead answer H-A-91 (3)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:39:29 +01:00
jschoubben 9fe6330d79 B4: the bridge's whole test says head-moved; bridge_heads reads the heads held
The bridge's whole test runs the program against an embedded bus as the node does: a start, a
merge, an admitted push to an unprotected branch and a restart after a move each give one
head-moved naming the head read through the seat git's branch and the head held before (start
true at a start); a head heard twice, a branch not watched and a restart at the head held give
none, and the bridge asks the seat git nothing but branch. Its bus permissions are composed for
the manifest as it now stands. The pull, open pull request, group and unprotected-branch tests
go; the owed test performs the gate's statuses once, after a restart and again by hand.

bridge_heads answers each key of the state heads with the head the bridge holds, read-only.

Builds: design 47 H §739 (how checked), §774; lead answers H-A-48, H-A-85 (1)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:34:07 +01:00
jschoubben 7ac48fd882 B2, B3: the forge bridge says head-moved and carries the one model
B2: on every branch-moved of a watched branch, and at its start for every watched branch's head,
the bridge reads the head through the seat git's verb branch, as a kept call that waits on the
holder's return while the branch is watched, and writes one head-moved (repository, branch,
head, before the head held in its state heads, start) on its stream forge; a head equal to the
one held writes nothing. It reads no pull request, no tree, no file and no protection: the
merged and pushed messages, their inputs and the state inputs, the call of git.pull, the finding
branch-unprotected with its held finding writer, and the ten-second redelivery go. A move whose
head cannot be read is left unacknowledged for the bus to deliver again.

B3: the pull-message path (Convert, the open pull requests at start and reconnect, the
forge-delete message), the groups (internal/groups' group part, forge.group.merged, end-asks,
bridge_groups, bridge_group_end_again), bridge_merged_message with the forge-made reader, the
checker's owed and the merge act of a switched key (with the switch's bookkeeping read) go, with
their consumes, invokes, states and requires. The gate's owed path stays (H-A-71), its catch-up
from the gate's verb groups gone with that verb.

The two meet in main.go and the package compiles only with both, as it did not since B1. The
unit suites pass; the whole test (integration) is rewritten in B4.

Builds: design 47 H §739, §764 (the forge bridge; the branch protection); lead answers H-A-23, H-A-27, H-A-48, H-A-71

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:31:04 +01:00
jschoubben 3560362330 C-W2: a step the assigner serves names its asker
The seat mesh-wants' StepRun gains asker, the seat name of the part asking the step (the runner
writes mesh-pipeline-runner), so the assigner names a step's progress with it. Optional in the
definition: the line names its writer, not that a step without it is refused. Both accept:step
fixtures carry it; the generated fixture test fails without the field.

Builds: design 47 H §745; lead answer H-A-88 (hq 2c6c7c03), H-A-41

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:27:41 +01:00
jschoubben 87e9084e15 C-R2: the runner's pipelines key a run on request; a mover keeps its heard publishes
The seat mesh-pipeline-runner's state pipelines gains the key form <branch>.run.<run>, one
pipeline per run on request, its movers the named commits' modules (it delivers nothing). Its
Mover gains published, the artifact references whose package-published (or, on a kind-3 edge,
whose build step's passed end) the runner heard at the pipeline's branch and source, kept with
the pipeline so a dependency's wait survives a restart. A fixture of the new key form with a
mover naming one heard publish proves both.

Builds: design 47 H §746; lead answers H-A-83 (1) (hq ae7fc40a), H-A-84 (hq main 0631f43f+)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:21:43 +01:00
jschoubben b75dcc947c C1b: the seat mesh-catalog keeps no state verdicts
With the validation gone from the take-in, the state verdicts, its value and the client's
helpers over it (VerdictsKey, TakeVerdict, VerdictWhy, ModuleVerdict, the codec) leave the
seat's contract; a module's own verdict field keeps its three states.

Builds: design 47 H §745, §764 (the indexer: the validation at the take-in); lead answer H-A-79

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:15:15 +01:00
jschoubben 630c099972 B1: the forge bridge's one message, head-moved
forgemsg holds one message per move of a watched branch's head: head-moved with repository,
branch, head, before and start, on forge.<owner>.<name>.<escaped branch>.head-moved, refusing
a head equal to the one held and any field it does not have (trees, base). The messages pull,
merged, pushed, refused and open, their trees, inputs, group and members, and the group's end
forge.group.merged go; owed.ended stays.

Builds: design 47 H §739, §764 (the forge bridge: merged, pushed, base, the group's end); lead answers H-A-27, H-A-48

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:13:15 +01:00
jschoubben f9fa0ec2b2 C-W: a want awaiting its hand placement is held
The assigner holds the want of a node whose node-engine does not state want:version in the
want's hold field, its words awaits-hand-placement; the client takes that hold beside a put's
and refuses a hold with other words.

Builds: design 47 H §772; lead answer H-A-73

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:09:23 +01:00
jschoubben 83d454217f steprun: the asker judgement goes; every asker's kind is taken
The step runner no longer reads the seat list for the callers a build seat's definition
lists, refuses an asker it does not name, or gives an ask back until the list says:
Puller.Callers, CallersFromDefinitions, SeatListCallers and the seat list reader's start
go, since no seat definition names callers and every account may call every verb. The
step is still keyed under the asker its subject names, and main's subject still takes
only a build, a run and the checker's check and recheck. mesh-delivery-sdk builds again
against mesh-sdk without seatdef callers.

Builds: design 47 H §752, §764 (every seat contract: the callers lists); lead answers H-A-41, H-A-66, H-A-71 (3)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:09:01 +01:00
jschoubben 4599d2e60f C-W: the seat mesh-wants: a want by version, pipeline and rank; pipeline-wanted; validate
A module want names its fingerprint and released version with the pipeline (<branch>.<source>)
and rank (its tier, from 1) it was written from, neither for a placement's build; parts, rule,
newest, kept-on and composed go, and the client checks a rank with its pipeline instead of a
composition and a rule. The node's own want is unchanged. The accepted step and its end take
validate beside composition. wanted gains pipeline; pipeline-wanted.<branch>.<source> says a
pipeline's wants once at their end, running or tier-late with the movers and machines not
running. version-delivered goes.

Builds: design 47 H §745, §750, §751, §764 (the assigner: version-delivered, the digests in a want); lead answers H-A-37, H-A-51, H-A-56, H-A-72 (6), H-A-75

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:05:15 +01:00
jschoubben b7b4087095 C-R: the runner's seat: one pipeline per take-in, its movers, the queue and place
The pipeline is keyed <branch>.<source> (and <branch>.<source>.published.<package>.<version>
for a publish), its movers {module, commit, fingerprint, repository, tier} carry what was
per module (version, widens, tests, over-budget); group goes, run, published, unlogged and
end-count stay; call and seq are added and the outcome gains waiting. Each step result
names its module and the step enum gains validate. pipeline-started names the movers;
pipeline-ended carries per mover its fingerprint, version, steps and tier. pipeline-started,
pipeline-ended, wait-held and wait-ended are said on <branch>.<source>; the step events and
run-started on <module>.<branch>.<commit>. A wait gains step pipeline and on
pipelines-at-once. The state queue holds queue.next and waiting.<sequence>; the verb place
takes pipeline and why. group-ended and its body go. The client builds and parses both
subject forms.

Builds: design 47 H §745, §746, §747, §748, §764 (the runner: group-ended, the key per module); lead answers H-A-35, H-A-36, H-A-51, H-A-52, H-A-53, H-A-72 (1) to (5)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 03:01:24 +01:00
jschoubben d24d6a0890 C2: a kept call's member names its base branch
So the forge module can say merge-called for what landed after a restart
without reading the forge again, each member of the state calls keeps the
pull request's base, read at judging.

Builds: design 47 H §766; lead answer H-A-71 (2)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:51:26 +01:00
jschoubben 5042d9c426 C-acts: the build seats' acting rules name no bound
acts stays as the declaration the graph's edges and tiers are built from; the
bound of the build seats' rule goes with the acting part's hold. The generated
contracts do not change.

Builds: design 47 H §750, §764 (the acting part's hold); lead answers H-A-57, H-A-70

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:51:01 +01:00
jschoubben 6e027de597 C3: the source revision's contract: source-moved per source, record-opened per call, heads
source-moved is said on source-moved.<branch>.<source> with v, branch, source,
call, start and heads [{repository, commit}], kept 30 days and 64 MiB on its
own stream; its wildcard has two tokens (H-A-26). record-opened is said on
record-opened.<call>; it and record-ended carry branch, source and call, the
milestone <branch>.<source>, no forge-made fields (H-A-25). The source value
carries source and call, the verb source answers call; the state heads keeps
each untaken head {head, since} by <repository>.<branch>. The states merges,
asks and turns, the verb groups, and the events approval-carried, pinned,
freed, pins-stated and heads-replaced go (H-A-28, H-A-48), with their helpers
in meshdeliveryclient.

Builds: design 47 H §737, §740, §749, §764 (the gate's states, groups and switch); lead answers H-A-23, H-A-25, H-A-26, H-A-28, H-A-33, H-A-48

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:49:32 +01:00
jschoubben f8bbde9627 C2: the merge call's contract: merge-pull takes pulls and by, merge-called, calls
merge-pull takes a list of one or more members {repository, number, head} and
by, judges no branch protection and answers once, at the call's end, the call
and each member's outcome {repository, number, state, commit|why} (H-A-21,
H-A-22); its caller id goes, and it names a timeout of 120 s (H-A-43). The
seat says merge-called on merge-called.<call>, kept 30 days and 64 MiB on
SEAT_git_merge-called, its body in bodies; the state calls keeps each call
with said (H-A-47). branch answers no protected.

Builds: design 47 H §734, §735, §736, §766, §767, §764 (the single-pull form, the protection judgement); lead answers H-A-21, H-A-22, H-A-23, H-A-24, H-A-43, H-A-47

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:44:23 +01:00
jschoubben f006ab50bc C1: the take-in's form is the lines' fields; policies gain validate
take-in.taken carries branch, source, revision, call, entries (one
{module, commit, fingerprint, repository} per mover, no pull), skipped, and by
and why for a hand act; take-in.pending carries branch, source and why; the
keys, kinds, merges, groups and the other fields of the earlier take-in go, so
readers pair a wait by <branch>.<source> (H-A-31). StepStatuses gains
validate before composition; the seat says no manifest-accepted or
manifest-refused.

Builds: design 47 H §743, §744, §745, §763 (G §476, §648), §764 (the indexer's manifest events); lead answer H-A-31

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:41:07 +01:00
jschoubben 8e5e5dbb42 seatgen: an identifier splits on ':' too, so want:version is a Go constant
The node-engine seat's enum value want:version must yield a parseable Go
constant; ':' splits as '-', '_', '.' and a space do. No existing generated
output changes (the generator's check of mesh-delivery-contracts is unchanged).

Builds: design 47 H lead answer H-A-64

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:36:37 +01:00
jschoubben ce6edcd9b9 C0: no seat definition names callers; the generator refuses one that does
Every account may call every verb (ADR 0682 §10), so the callers lists of every
seat definition and part of mesh-delivery-contracts go, and the generator, the
one reader of definitions in this repository, refuses a verb or accepted kind
that names callers instead of checking its forms. The generated contracts do
not change: the generator made nothing of callers.

Builds: design 47 H §752, §764 (every seat contract: the callers lists)

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-20 02:36:30 +01:00
jschoubben bf642c45b7 mesh-delivery: the old hand-kept mesh-catalog contract copy loses the config-candidates layout (issue 1982; ADR 0681 §2)
The copy under contracts/go/seats/meshcatalog still held the layout, its key helpers, its constant and two fixture cases; nothing
in contracts/ or sdk/ uses them, and ADR 0681 §7's check greps the trunk for the name.

Builds: ADR 0681 §2 (mesh-sdk keeps no config-candidate layout; the contract keeps none), §7.
2026-11-19 23:05:21 +01:00
jschoubben 64ae24aef7 mesh-delivery: remove the indexer's state config-candidates and its readers (issue 1982; ADR 0681 §2)
The mesh knows no pull request (design 47 G, G-1 §38-39), so the indexer keeps no pull request's config declarations.
Removed: the code that writes the state (control_records.go: the candidate writes after a merge and for a pull
message), its bucket, view family (views.config-candidates) and read tool (indexer_read_config_candidates) in the
indexer's module.json, the view's tool indexer_config_candidates and its read of the indexer's tool, the state and its
fixtures in the seat mesh-catalog's definition (generated Go and TypeScript regenerated by seatgen). indexer_config_readers
no longer names an open pull request as a reader of a key. The state `predefined` stays.

Builds: ADR 0681 §2 (the indexer keeps no config candidate: state, read tool and view, the seat contract's layout; `predefined` stays), §1 (indexer_config_readers names no pull request).

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 22:47:44 +01:00
mesh-tools 0ea4b60dbd Merge pull request 'steprun: only runs that started count toward poison; every end of an ask is written, said and acknowledged (A-489, A-489a, A-489b, A-489c)' (#257) from fix/build-node-step-ends into main go/v0.1.73 2026-11-19 21:40:15 +00:00
mesh-tools 5da42d1775 Merge pull request 'Catalog versions: each module's version floor from its Go tags and its package holders, per module, as a kept call (A-490c..A-490i, §412)' (#260) from fix/catalog-versions-first-fill into main 2026-11-19 21:40:12 +00:00
mesh-tools 0639ec4bca Merge pull request 'Builder: build again a done trunk key whose version is unreleased, reserved afresh; a trunk build without its version fails (A-490, A-490b; carries the reservation takeover A-455..A-473)' (#259) from fix/builder-unreleased-build-again into main 2026-11-19 21:40:10 +00:00
mesh-tools ac833a48f8 Merge pull request 'publishconformance: the cases every package holder is held to for a publish with a digest and no parts (A-490, A-490a)' (#258) from fix/gitea-publish-by-digest into main 2026-11-19 21:40:02 +00:00
jschoubben d9596548f9 publishconformance: a version held under another digest is refused though reserved with no digest yet (A-490j, A-490a)
Builds: design 47 G A-490j, A-490a

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 21:39:21 +01:00
jschoubben 9a7cd5861d Builder: a published kept at a version whose recorded digests none matches raises the finding at once, the version taken and never released (A-490j)
Builds: design 47 G A-490j

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 21:34:58 +01:00
jschoubben 95b4b308ad publishconformance: the cases of a one-part publish made while the build runs, reserved with no digest yet among them (A-490j)
Builds: design 47 G A-490j, A-490a

Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 21:33:46 +01:00
jschoubben b413949ea8 Catalog versions: the floor's finding clears on every exit of the fill, and the integration test shows the kept call asked again
FillFloor clears the module's call-unanswered finding and its call-since key on every exit but the end of its context (a release
gave the module a version, it left the trunk, the callee answered with a failure), as ADR 0668 §3 has: FloorModules never
picks the module again. The module's clock starts at its first unanswered try, not when the form gives up. The integration test
sets the form's bound to one second, waits for the module's call-since key in the state heard, and only then serves git.tags;
it fails with `return err` at the unanswered try. Unit cases for the three exits.

Builds: design 47 G §412, A-490g, A-490h, A-490i
Claude-Session: https://claude.ai/code/session_01JG73KHUnBze48rdua8u2sy
2026-11-19 17:42:56 +01:00