Gate: final tip 27759a1b vs a8c5fbf9 (10.4.10) control — collected
3,212, 0 new reds after two fix cycles (coverage-guard registration +
perf-lane classification; a real budget flake in the batch-size test
switched to unvectored items). shasum ffc33df95b2709dfcc8c67ac961991e3153f8883.
The fleet's releases wall reads one public URL per product. These files are
that door for Brainy and Open Brainy: newest first, honest history from the
changelog, one entry appended by every release from here on.
The one-doc-set ruling (2026-08-31) gives soulcraft.com/docs to the paid
product alone; the site serves redirects for the slugs this rail used to
push. The push script stays in the tree as history; the rail stops calling
it.
Diagnosis of the "packed history is damaged" narration that fires on every
run of the affected stores. It is a WRITER defect, and the reader's refusal
was the symptom rather than the cause.
A sealed segment declares one contiguous range [firstGeneration,
lastGeneration], and every reader treats that range as containment:
coveringSegment is an interval test, hasGeneration returns true for anything
inside it, and open() seeds committedRanges from it.
repackHistory handed fold() a SPARSE batch. Three filters punch holes in its
candidate list mid-run — a generation absent from committedRanges never
appears, one still in the pending buffer is skipped, one whose tx.json will
not read is skipped — and fold() then computed the range from the first and
last survivor, claiming every generation in between. The next open merged
that mis-declared range back into committedRanges, re-admitting the hole as
committed history, so the following auto-compaction pass asked the packed
tier for a frame that was never written and failed. Re-merged at every open,
which is why it repeated on every run.
Confirmed against a forensic fixture: generation directories 1..2503 present
except exactly one, 1416; and its fact-log segment already showed the tell —
seg-...1410.bfl declaring 1410..1940 (531 generations) while recording 530
facts.
Three changes:
- repackHistory folds each contiguous RUN as its own segment
(`contiguousRuns`), so ranges describe exactly what the segments contain.
- fold() REFUSES a non-contiguous batch, naming the gap and its width. The
density law is now mechanical, so no future caller can reintroduce it. A
refusal loses nothing: the generations stay live and readable.
- Stores already carrying the damage heal instead of wedging. A segment
whose declared span exceeds its frame count is SPARSE; `actualRanges()`
reads the real generation list from its sidecar so open() never re-admits
the holes, and readFrame reports such a hole as unpacked with a narration
naming the segment, rather than throwing. A DENSE segment missing a frame
is still loud damage — that one means the manifest and sidecar disagree.
Pins: nine unit cases (refusal and its message, honest ranges for separately
folded runs, a reconstructed pre-fix sparse segment serving its real frames
while reporting holes as unpacked, holes excluded from actualRanges, and the
dense-segment damage path still throwing) plus an end-to-end case that
deletes a generation directory and drives the real sequence — ordinary
close()-time repacking folds over the hole, then reopen and compact must both
complete. Verified red without the fix: the segment declared an
11-generation span while holding 10 frames.
Two halves of one defect, found by a seeded-SIGKILL crash lane.
THE FALSE POSITIVE. stampEntityTree() recorded generationStore.generation()
— the ALLOCATED counter, a number a write in flight has claimed and may
never commit — while the JSDoc beside it already said the source is the
committed generation. Every crash inside a write window therefore produced
a spurious verdict at the next open: either 'sourceGeneration N is ahead of
the log head N-1' (the allocated generation died with the process) or
'rollup invariant nounCount: stamped X, observed Y' (the recovery fold
folded facts the stamp's counts predate). Both told the operator to run
repairIndex() — a whole-store recount — for a store that was coherent.
Measured before this commit: 4 of 11 SIGKILL cycles on a healthy store
raised one of the two. The stamp and the open now both read
committedGeneration(), which is what every other open-time watermark in the
class already reasons about.
THE TERMINAL VERDICT. A stamp still ahead of committed truth after the
recovery fold witnesses a generation that is not in the log — the stamp's
fsync outlived the tail's, and there is nothing to arrive. That is its own
verdict state now ('torn'), never folded in with 'incoherent': the two have
opposite cures. A writer open demotes it — the unusable stamped surface is
re-derived at the committed generation from the live counters, O(1),
straight-line, no loop and no await on external progress, narrated with
both count sets, the stamp's path and its committedAt. A read-only open
cannot re-stamp, so it says so and names the cure instead of guessing, and
still serves. Neither branch waits, and neither locks an owner out of a
canonical tree the stamp only describes.
Pins: the verifier returns the torn verdict with both generations; a
fabricated head-behind-source store narrates precisely, demotes inside a
bounded open, serves its rows, and is quiet at the next open (the demotion
converges); a read-only open narrates the same verdict and leaves the bytes
untouched.
Step 11 POSTed to repos/soulcraft/brainy while printing the correct URL; dormant only because FORGEJO_RELEASE_TOKEN was unset. Found during the 10.4.4 cut verification.
The release gate's own output caught this: every test brain printed
"[VFS] old-root sweep complete in 1ms and recorded" — hundreds of lines — and
a consumer would get two of them on the first open of every store.
They were emitted on the always-visible channel, which a production log level
deliberately CANNOT silence. That channel exists so an operator can always
learn why a database is slow; a 0ms no-op on a fresh store is not that, and
announcing it there trains people to ignore the one channel built to be
impossible to ignore. It was also inconsistent with every other narration in
this work, all of which is silent under a threshold.
The sweep now speaks when it has something to say — duplicate roots removed, or
a wall over a second that a person watching a slow first open deserves
explained — and otherwise does its work, records its marker, and stays quiet.
cleanupOldRoots() reports what it removed so the decision rests on a fact
rather than on a guess.
Pin: a fresh store's sweep emits nothing on the channel and still records its
marker, so the silence can never be mistaken for the work being skipped.
Two gate failures on main, one real and one long-hidden.
THE HEALTH-GATE PIN encoded the old law — "narrates once per generation, twice
across a generation bump" — which the content-keyed dedupe deliberately
replaced. A provider's `generation` bumps on every ledger mutation and every
rebuild boundary, so keying narration on it re-printed an unchanged health line
on every read that consulted a busy provider, and let a provider that never
bumped suppress a line whose reasons had genuinely changed. The pin now asserts
BOTH directions: an unchanged verdict stays silent however the counter moves,
and a changed verdict is always heard.
THE VFS HYBRID-SEARCH SUITE configured its store with `options.basePath`, an
alias removed at the 8.0 major that configures nothing. The suite was therefore
never using its temp directory — it opened the DEFAULT store, shared with every
other run on the machine, and accumulated tens of thousands of rows until it
failed on that shared store's graph adjacency instead of on anything it tests.
It now passes `storage.path`. The suite drops from 6.5s to 0.3s, which is the
measure of how much foreign data it had been opening.
Neither failure was caused by the release branch; the first is the branch's own
behaviour change meeting its outdated pin, the second predates it.
Correctness and observability for 10.4.4: the writer-lock clean-close record
and the always-run terminal releases, the self-healing count ledger and an
atomic counts.json, always-visible open and repair narration, the non-blocking
open for a provider rebuilding itself, an event-driven flush-request watch, and
the operator-visible operator set (three served, four refused by name).
The manifest's prose pointer named a document that answers a confidential
specification, and such a document does not belong in a public repository even
in summary. The pointer is dropped — the manifest is generated from this
engine's own surface and is self-describing — and the requirement marking it
deliberately omits is recorded with the contract's owner rather than here.
The standard is written down so this is not relitigated per document.
A release audit found hostnames, store identities and operational anecdotes in
this branch's commit messages — not trade secrets, but nothing a public
repository's permanent history should carry either. The messages were rewritten
to keep every number and drop every provenance; the rule is written down here
so the next measurement does not have to be caught by an audit.
MEASURED on a 14,056-noun / 72,679-verb production-shaped store, measured
solo under an exclusive lock: the vfs-bootstrap
phase costs 37.8s on main and 38.0s on this branch — unchanged — and NO
"vfs.init" step line was emitted at all, meaning the VFS's own init fell under
the 2s narration threshold. The phase is therefore almost entirely NOT the VFS,
and the old-root sweep this branch moved to the background was never what made
it expensive.
What else lives in that span is now named: the log-authority artifact read, the
adoption ORACLE (which verifies the log against canonical before flipping a
brain to durable-at-ack), the legacy pending-embed sidecar bridge, and the
pending-embed recovery fold. One of those holds ~38 seconds of every open of
this store and the next measurement will say which, by name, instead of leaving
a phase label to be guessed at.
The safety sweep is armed alongside the watch, and startFlushRequestPolling()
declines to arm over an existing interval — so when a watch died mid-life the
fallback did nothing and the store quietly answered flush requests on a 30s
cadence instead of the 500ms one the door promises. The sweep is cleared first.
A degrade nobody asked for is still a degrade.
Arming is asynchronous — the request directory is created before it can be
watched — so during that window neither the watcher nor the sweep interval
exists yet and the guard let a second call through, leaving two watchers and
two sweeps for the life of the store. The callback is the flag that covers the
window.
Three idle-burn items from the steady-state audit, and one correction.
THE FLUSH-REQUEST WATCH (the strongest of them). It readdir'd the request
directory every 500 ms, per brain, for the life of every writer — armed on
every non-reader brain whether or not any inspector process existed. In a process holding many stores that is tens of directory reads per second
on a completely idle service, plus a stale-request GC on every one of them. It now
uses fs.watch, so the arrival itself wakes it and a request is seen SOONER
than the poll saw it. Two concessions ride along, both stated in the code: a
30s safety sweep (fs.watch drops events on some network and fuse filesystems,
and the GC needs a tick of its own — two orders of magnitude fewer reads than
the poll made), and a fall back to the original 500 ms poll, narrated, on a
filesystem that cannot watch at all, because an inspector whose request is
never seen waits forever.
THE WRITER HEARTBEAT goes 10s → 60s. It is observability ONLY — staleness is
decided by pid liveness and the fence compares pid + hostname, so no decision
anywhere reads the timestamp — and at 10s it was a lock-file write every ten
seconds per brain forever, for a value nothing computes with. An operator
still sees a heartbeat inside the minute.
THE HEALTH NARRATION dedupes by CONTENT, not by the provider's generation
counter. That counter bumps on every ledger mutation and rebuild boundary, so
a provider bumping it on routine work re-emitted the same unchanged line on
every read, while one that never bumped could suppress a line whose reasons
had genuinely changed. The generation is still reported; it no longer decides
whether the line is worth saying.
CORRECTION, and it is against my own earlier claim: the idle-flush commit read
a reported idle-CPU observation (many stores, no writes, a flush every ~35s,
over a core burned) as caused by
the flush path. That does not follow — this engine's cadence is write-driven
(every trigger runs through noteWriteForPersistence, which only a committed
write calls), so something was CALLING flush() on those brains and the caller
is still unidentified. The clean-flush gate makes such a call free; it does not
account for it. The code comments and the idle lane now say exactly that.
Pins: tests/integration/flush-watcher-event-driven.test.ts — an idle writer
makes at most one request-directory read in 8 seconds (the old poll made ~16),
and a dropped request is still acked well inside the safety sweep.
The 7.x-to-8.0 layout probe runs on the open path of every store that does not
yet carry its completion marker — a restore, a store built by an older release
— and asked whether the canonical tree holds anything by LISTING it: a
recursive walk of every file in every entity directory, to learn a boolean. It
now asks the one-level door added for generation discovery, falling back to the
listing on an adapter that lacks it.
Also files a defect found while ratifying the operator set: the VFS builds its
path-prefix filter as `$startsWith`, an operator no engine spelling accepts,
so vfs.searchFiles({ path }) throws INVALID_QUERY on every call that passes a
path. Pre-existing, unrelated to the operator work, and left as a filing —
a path-prefix search needs a design answer, not a spelling correction.
MEASURED on a production-shaped store (14,056 nouns / 72,679 verbs, an 11 GB
generation history), measured solo under an exclusive lock: the generation-store phase cost 55,538 ms of a WARM
REOPEN after a clean close — with the fold correctly skipped, so nothing in
that phase's name explained it.
This is what it was doing. Discovering which generations exist on disk called
listRawObjects('_generations'), which RECURSES the whole tree and returns
every file in every generation directory — to extract a set of integers that
the top-level directory NAMES already spell out. The cost scales with the
entire history, is paid on every open, warm or cold, and grows for the life of
the store.
A one-level door — listRawPrefixes(prefix), the immediate child directory
names — is added to the storage seam. The filesystem adapter answers it with a
single readdir; BaseStorage derives it from the recursive listing, so an
adapter without a cheap implementation is never wrong, only never faster; and
the generation store falls back to the old listing when the door is absent.
One behavioural difference, stated: an EMPTY generation directory is now
discovered where the file listing could not see it. Above the committed
watermark that is a crash scar, and recovery already has an explicit branch
for it ("indeterminate partial dir" — dropped, narrated). Below it, it becomes
a resolvable generation holding no records, which is what an empty generation
means.
Suites: the durability kill matrix (15), db-mvcc (30), history repacking (4),
rollback trapdoor (3), entity-tree stamp (4) and the full unit suite (2,105)
all green.
Both mutate durable state outside the two commit paths, so neither was seen by
the flush witness added with the idle-flush law. A clear() followed by a
flush() would have found the brain "clean" and skipped the entity-tree stamp,
leaving a stamp describing the population the clear had just removed — a false
divergence warning at the next open. Closing the gap where it is, rather than
widening the witness to guess.
A phase that costs a minute and names only itself tells an operator where to
look but not what to look at. MEASURED on a real 14,056-noun / 72,679-verb
store, the warm reopen's generation-store phase cost 55,538 ms with nothing
inside it named — the fold was skipped (the close was clean), so the cost was
somewhere else entirely and the breakdown could not say where.
Six steps inside the open now report their own wall with their own cause when
they exceed the phase threshold: the generation store's open (manifest,
committed ranges, fact log, packed tier, crash replay), the entity-tree stamp
verification, the brain-format read, the pre-upgrade backup, the derived-index
gate, and the VFS init. Silent under the threshold, so a fast open says nothing
extra. Same always-visible channel as the phase lines.
MEASURED on a 14,056-noun / 72,679-verb production-shaped store, measured
solo under an exclusive lock: the vfs-bootstrap phase cost 43,021 ms of a cold open and 52,696 ms of
a WARM REOPEN. What dominates it is a migration sweep — a filtered find() over
the whole store hunting for root directories created before the fixed root id
existed. A store either carries such duplicates or never will, and the sweep
ran on every open, forever, in the foreground.
It is now caused by the store's state instead of by the open count: a durable
marker under _system/ records that the sweep has run, and a store carrying it
never sweeps again. A store without one sweeps in the BACKGROUND — the sweep
only removes duplicate roots, nothing serves from them, and it was already
declared non-critical — narrated at both ends, with whenRootSweepSettled() for
anyone who needs to observe rather than race it. An adapter with no raw-object
door keeps the old behaviour: correctness over cost, never a silent skip.
Pins: tests/integration/vfs-root-sweep-once.test.ts — the sweep runs on the
first open and never on the second or third; a sweep slowed to 4s does not
delay the open.
Open Brainy's side of the API contract the accelerated engine published.
DECLARED: package.json carries "brainyContract": 1 and the engine states its
own via contractVersion() / BRAINY_CONTRACT_VERSION — two engines compare an
integer instead of probing prototypes, and a tool reads the package field
without importing the engine. Pinned so the two can never drift apart.
SERVED: hasAll, noneOf and excludes now work on the index path. The defect
underneath was worse than the reported divergence — the metadata index's
operator switch had NO DEFAULT CASE, so any operator without a case left the
field's match set at its initial [] and find() returned an empty page.
Documented, validator-accepted, matcher-implemented operators answering
silently wrong. hasAll intersects each element's posting set (an empty operand
is vacuously true of every row that has the field), noneOf complements their
union, excludes complements contains.
REFUSED BY NAME: startsWith, endsWith, matches and length raise
INVALID_QUERY naming the operator, the field and the reason. An equality/range
posting index cannot evaluate a substring, a pattern or an array length without
reading every row — which is the cost this path exists to avoid — so it refuses
rather than answering an empty page. Both engines now agree on all 25 tokens
and contract 1 has no remaining operator divergence. This is a visible change
for a consumer calling those four through find({ where }): an empty page
becomes a typed refusal.
EMITTED: scripts/emit-contract-manifest.mjs generates docs/api-contract.json
from the BUILT surface — prototype doors, exported error classes, the operator
sets read out of their single definitions, the field-addressing vocabulary, the
health verdicts. Nothing hand-maintained, so a diff between two manifests is a
diff between two engines. `--check` fails on a stale manifest, which makes the
announce-every-addition duty mechanical rather than remembered.
RATIFIED in docs/contract-1-ratification.md: the 41-of-57 required split with
the promise spelled out (a refusal is part of a door; deprecation is not
removal), the serving-withholding list confirmed exhaustive and identical, the
minor/major rule adopted with the announcement duty, the 30 storage seam
methods committed as supported surface until Stage 2, and a finding filed
against the spec — is / isNot / greaterEqual / lessEqual are listed there as
served aliases and have never existed in this engine, which throws
INVALID_QUERY on all four.
Follower to the self-rebuild deference. The open gate's consistency check —
"metadata index has 0 entries but storage has N entities" → CRITICAL + a forced
second rebuild — knew two states, migrating and not. A provider whose rebuild()
returns once the rebuild is OWNED AND RUNNING ONLINE (its doors refusing by
name while other families serve) legitimately reports 0 entries there, so every
first contact printed a false CRITICAL and kicked a redundant second rebuild.
The exemption rides the rebuild-progress hook, NOT isMigrating() — widening
that would hold every write and 503 the whole brain through the migration
snapshot, which is worse than the false alarm. The check's real class is
untouched: a provider reporting 0 entries with no rebuild in progress still
trips it.
The crash-recovery rebuild kick gets the same deference: a provider already
rebuilding itself from canonical is doing exactly that work, and the fold ran
in the generation store's open before any provider existed, so what it is
reading is the repaired canonical.
Pin: a provider stub reporting a rebuild and 0 entries opens with no CRITICAL
line and no second rebuild; the vacuous-stub case fails loudly.
MEASURED on a production store: a metadata provider that had to rebuild made
init() pay the ENTIRE rebuild on the foreground — 641 seconds — with every
other family idle behind it. The cause is a missing distinction: a provider
reporting serving:false because it is BUSY BUILDING ITSELF and one reporting
serving:false because it is BROKEN looked identical through healthReport(),
and both were answered the same way — call rebuild(), and wait for it.
The contract that tells them apart is one optional, synchronous, O(1) hook:
`rebuildInProgress(): ProviderRebuildProgress | null`, reporting a phase name
and whatever the provider actually measures (done/total/startedAt) — never an
estimate dressed as a fact. A provider without the hook behaves exactly as
before.
With it, a provider owns its own rebuild:
- the open gate neither starts a second rebuild nor waits for the provider's,
and narrates that it is not waiting and what will refuse meanwhile;
- init() returns and every other family serves;
- that family's doors refuse BY NAME, carrying the provider's own progress,
and say plainly that the door opens by itself and no action is needed —
distinct from a broken index, which names repairIndex();
- the epoch stamp does not advance while any family is still being built.
Nothing is ever served empty: a not-serving family refuses, as it already did.
Pins: tests/integration/open-does-not-wait-for-a-rebuilding-provider.test.ts —
init() returns in milliseconds against a provider claiming a 6s rebuild, brainy
starts no rebuild of its own, a filtered read refuses naming the phase and the
4,096/14,056 progress, and the door answers once the provider reports serving.
The pin fails loudly rather than vacuously if its stub never installs.
REPORTED from the field: a process holding many stores, with no writes for ten
minutes, printed "All indexes flushed to disk in 216-601ms" per store every
~35 seconds and burned over a core at idle. Every one of those flushes re-persisted
state identical to what was already on disk — the provider flushes, the
watermark stamps, the generation counter, the entity-tree stamp — because
flush() never asked whether anything had changed.
- flush() over a clean brain is now O(1) and silent: a dirty witness is set by
every committed write (both commit paths end at noteWriteForPersistence, and
the deferred-embed worker lands through the single-op path) and cleared by a
flush that runs. A write landing DURING a flush sets it again, so no write's
work is ever skipped — it is done by the next flush. Set before the policy
check, so a `'manual'` consumer's explicit flush is never a no-op it didn't
ask for.
- An explicit flush now tells the cadence it happened. It didn't, so the very
next write saw "30s since the last flush" and kicked a background flush with
nothing to do, and the idle timer fired two seconds later over writes the
explicit flush had already persisted.
- The graph adjacency index's auto-flush asks before it acts: two O(1) reads
of the LSM MemTables, and a tick over a quiet index returns without calling
into the trees at all.
assessProviderHealth is NOT timer-driven — it is a synchronous O(1) read of a
provider's own healthReport(), called on the read gate, so it costs nothing on
an idle brain. No change needed there.
Pins: tests/integration/idle-costs-nothing.test.ts — 90 idle seconds produce
zero flushes, zero provider calls and zero log lines; three explicit flushes
over a clean brain call no provider; one write earns exactly one flush.
On a production store (14,647 nouns / 73,070 verbs) a repairIndex() ran for
more than thirty minutes at roughly a full core with ZERO log lines between
its start and its end, while the read doors kept serving. The operator could
tell it was alive only from `top`, and could not tell which of its
single-threaded walks it was inside.
Same law as the open, applied to the repair:
- every phase announces itself BEFORE it works, naming what it is about to
walk (each canonical walk, the VFS containment reconciliation, each
provider's invariant pass);
- an unref'd heartbeat names the phase still running every 5s, for as long as
it runs;
- every phase reports its own wall, and that wall is carried in the TYPED
receipt as RepairFamilyReport.durationMs — a receipt that cannot say where
the time went is not a receipt;
- the whole repair's narration moves to the always-visible channel, so a
production log level cannot silence it.
The phases move into runRepairIndexPhases() so the heartbeat can live in a
finally around them; the public door and its report shape are unchanged apart
from the added durationMs.
Pins: tests/integration/repair-narration.test.ts — every checked family has a
start line, a finish line with its wall, and a numeric durationMs in the
receipt; a phase slowed to 6.5s produces a heartbeat naming it, with the
logger clamped to ERROR.
MEASURED on a real store: the ALL-visibility ledger read 14,231 nouns against
14,056 identity records and 72,729 verbs against 72,679 — exactly that store's
25 noun and 50 verb SCAR directories. Two copies of the same archive derived
different numbers (14,231 and 14,081), because each had been persisted at a
different moment under the old rule that counted one entity per id DIRECTORY.
A downstream index heal subtracted against that denominator and reported
remaining work that did not exist.
The scan already applies the right predicate — one entity per IDENTITY RECORD
(the metadata content leg), shared with pruneOrphanedEntities so the two agree
by construction. What was missing is that a ledger persisted under the old rule
was only FLAGGED suspect and then went on serving its wrong numbers for the
life of the store, waiting for an operator to run repairIndex.
- The ledger now derives itself honestly in the BACKGROUND after the open,
narrating start and finish with the correction it made. Background because
these scalars are denominators — no read is served from them — and because
walks exactly like these are how a 24,898-id store spent minutes of a
restart in silence. Observable via whenCountLedgerSettled(); nothing in the
read path waits on it.
- A derivation that raced a write refuses to stamp its number "exact": one
retry on a quiet store, then the ledger stays SUSPECT and says so, naming
repairIndex as the door that recounts under a barrier.
- The one derivation that CANNOT leave the foreground says why it cannot:
getNounCount()/getVerbCount() are served from it, and a background walk
would make a populated store answer "0 entities" — a wrong answer, not a
slow one. It narrates its start and its wall instead.
- counts.json is written temp+rename. A truncating write left a window —
measured at roughly 750ms after a flush or close — in which a concurrent
reader saw the file EMPTY; an unparseable ledger sends the next open down
the full-rescan path, so the cheapest file in the store was buying the most
expensive recovery.
- The writer lock's clean-close record is now consulted before the
same-process branch too: a restart reported "Re-acquiring writer lock ...
this is a bug" immediately after a clean close, sending an operator after a
leak that did not exist.
Pins: tests/integration/count-ledger-identity-record.test.ts (background
correction with scar and ghost fixtures, two copies of one archive agreeing,
counts.json never observed unparseable across 40 persists);
tests/integration/ledger-derivation-identity.test.ts updated to the new law —
the OPEN still never walks (proved by slowing the walk 1.2s and timing the
open), and the ledger heals behind it.
An operator watched a production service open a 16 GB store and print nothing
for three minutes before its first line of work. Two defects, both fixed here.
The narration was written to `prodLog.warn`, and every environment that looks
like production clamps the logger to ERROR — so the phase breakdown that would
have named the slow phase was composed and thrown away. `prodLog.narrate` is
always visible, like `error`: it carries the two things an operator is
entitled to hear from a database regardless of a cost setting — why it is slow
and what it is doing about it. `silent: true` still silences it; that is a
request, not a default.
And nothing spoke DURING a phase, only after the whole open. init() now runs
an unref'd heartbeat that every 5s names the phase currently running, its
elapsed wall and what it is paying for, plus one line per phase as it ends for
any phase over 2s. The generation-log fold's own progress and completion lines
move to the same channel and now carry their wall — they were invisible in
production, which is how an operator came to restart a converging fold three
times.
Pins: tests/integration/open-narration.test.ts — narrate() survives the clamp
that silences warn(); a 6.5s storage-init produces a heartbeat naming the
phase and a completion line naming its wall, with the logger clamped to ERROR.
A production restart made this necessary: a service stopped with exit code 0,
having awaited close() on every pooled brain, and its next boot announced
"Overwriting stale writer lock ... appears dead" for every store it owned.
Nothing had crashed. "The recorded pid is gone" is equally true of an orderly
restart and of a crash, so the verdict could not tell an operator which one
they had — and when the OS recycles a pid it fails the other way, refusing to
open a store whose writer died days ago.
Three changes, all at the law:
- close() is two parts, and the second is unconditional. The durable steps
(flush, markers, component close, plugin deactivate, buffer drain) move to
closeDurableSteps(); the terminal releases — the flush-request watcher, the
WRITER LOCK, the VFS timers, the terminal `closed` flag — always run. The
original failure is narrated with what it costs the next open, then rethrown.
- releaseWriterLock() writes a CLEAN-CLOSE RECORD (`locks/_writer.close`)
naming the lock generation it released; the next claim consumes it, so a
record can never vouch for a later crash. An open reads the record instead
of guessing: recorded → nothing to recover; absent → say so, and name the
crash recovery this open will now run.
- The signal path stops failing in a batch. It was one try around a loop over
every open brain, so the first instance whose flush rejected stranded every
remaining brain's lock and markers — at exit code 0. Now: per-instance
isolation, the generation store's close (the clean-shutdown marker, without
which the next open folds the whole log) is part of shutdown, the lock is
given up in a finally, and the handler no longer calls process.exit() when
the host application has its own signal handler — that race truncated the
host's own close() mid-flight.
Pins: tests/integration/writer-lock-clean-close.test.ts — completed close
leaves no lock and a consumed-once record with a silent reopen; a failing
durable step still releases and still rethrows; SIGKILL leaves the lock with
no record and the reopen names the crash; a host SIGTERM handler runs to
completion.
Branch plan (10 lines):
1. writer lock: clean-close record + always-release close [this commit]
2. open narration: an always-on channel; production clamps prodLog to ERROR,
which is why a three-minute open printed nothing
3. open narration: per-phase lines as each phase ENDS, with progress cadence
4. measure both real-store fixtures on the box, before/after
5. move the generation-log fold out of the foreground where the serving law
allows; durable resumable progress marker
6. same for the VFS bootstrap
7. counts: a legacy container-rule ledger must not keep serving wrong
denominators; counts.json written atomically
8. counts pin with scar directories; two copies of one archive agree
9. docs/canonical-layout-ratification.md — 12 facts confirmed/corrected
10. report: MEASURED before/after, findings, and whether this is 10.4.4
Prepares the repo for its new home at soulcraftlabs/open-brainy ahead
of the Forgejo transfer: package name, publish registry, release
script, and every install/import reference across docs, src, tests,
examples, and integrations now point at @soulcraftlabs/brainy on
The Source. The npmjs storefront leg and byte-identity pair
verification are stripped from the release script — The Source is
now the only publish target. README gains an Open Brainy explainer
and a registry note for consumers.
@soulcraft/brainy 10.4.2 was the last release under the old name.
The engine-pair seam law: a zero-norm vector never crosses an engine
boundary. The index belt already refused to insert one, but the canonical
write and the vectored-noun ledger still counted it, so a near-empty store
whose only vectored row was zero-norm read "1 canonical vectored vs 0
indexed" and threw a not-ready error at open, and a store's own legacy
zero-norm VFS root could trip the same gate before its VFS-init-time cure
ever ran.
- add()/update() (single and transact()) now normalize an explicit
real all-zero vector to the unvectored [] shape before the dimension
pin, the ledger flag, and the index ops ever see it (loud, one warn per
write, canonical write still succeeds).
- The legacy counts.json derivation walk (scanVectoredNounCount) excludes
a persisted zero-norm row, matching the live ledger's definition.
- A legacy zero-norm VFS root now migrates at open, before the vector-leg
gate evaluates, via one O(1) fixed-path read (torn-tolerant — skips
rather than aborting init on a torn root, letting the recovery walk
heal it) — independent of whether a VirtualFileSystem is ever
constructed this session.
- update({ id, vector: [] }) (and the same op inside transact()) is now
the sanctioned, idempotent unvector door: index removal, exactly-once
ledger decrement, no re-embed, and it clears a pending deferred-embed
marker rather than leaving it to re-vectorize the row later. The
combination with deferEmbedding is a typed refusal.
- JsHnswVectorIndex.rebuild() now skips a zero-norm/empty persisted
vector when repopulating from canonical (the same belt the live
add/replace paths already had), and health()'s index-parity check now
compares HNSW size against the vectored-noun ledger rather than the
raw metadata-entry count, since a store's VFS root is permanently
unvectored by design.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A canonical row persisted with vector: [] (a system row, a deferred embed
not yet landed, or any other legitimately-unvectored record) is a normal,
enumerable row -- but rebuild()'s storage walk had no guard against it.
storage.getVectorIndexData() derives its answer from the row's own record,
so it returns non-null for any existing noun whether or not that noun was
ever actually indexed -- rebuild() admitted such rows into the live graph
with a length-0 vector. A vector-less node could become the entry point (or
occupy any graph position); the next real insert then ran a distance
calculation against it and blew up with a dimension mismatch.
Fix at two layers in src/hnsw/hnswIndex.ts:
- rebuild() now skips any row whose vector.length === 0 before it ever
becomes a graph node (one summary count line, never per-row spam), and
restores the pinned dimension from the first real vector it loads --
previously the pin stayed null across a restart, since addItem/updateItem
are the only sites that set it and rebuild() never goes through either.
- addItem/updateItem now refuse a length-0 vector with a typed
EmptyVectorIndexError instead of ever pinning dimension to 0 or storing a
vector-less node, so no future fill/rebuild/load path can poison the index
silently. getVectorSafe's lazy-load "not found" check also missed that an
empty array is truthy -- tightened to catch it.
IndexOperations.ts's ReplaceInVectorIndexOperation rollback paths now skip
re-adding an oldVector of length 0 (never a legal index member) instead of
attempting an illegal empty re-insert on rollback.
biography.test.ts's final ledger-exactness assertion assumed every noun the
lane creates is vectored, including the VFS root counted in
vfsBaselineNouns -- but the root is deliberately persisted unvectored.
Corrected the expected formula to exclude it.
Adds tests/integration/index-skips-unvectored.test.ts pinning: rebuild()
indexes only vectored rows with the dimension pinned correctly; clear()
then real adds never trip a dimension mismatch; addItem/updateItem refuse a
length-0 vector; and a crash/repair cycle stays dimension-consistent.
The noun/verb pagination walks (getNounsWithPagination,
getNounIdsWithPagination, getVerbsWithPagination) listed shard contents by
filtering for vectors.json, while the canonical count ledger has always
counted a row by its metadata.json presence alone. A row with metadata and
no vector file was therefore counted by the ledger but never yielded by the
walk — a permanent "counted but invisible" phantom for any downstream
consumer that iterates the walk to account for the ledger's total.
Nouns now enumerate by metadata.json and hydrate the vector leg optionally,
yielding the sanctioned unvectored shape (vector: []) when it's absent.
Verbs enumerate the same way, but a metadata-only verb row can only be fully
reconstructed when sourceId/targetId happen to be recoverable from metadata
(never true for a current production write — those fields live only in the
vector leg); otherwise the row is counted but loudly skipped rather than
fabricated, since a phantom edge with fake endpoints would be worse than the
original defect.
Separately, GenerationStore's recovery-fold replay (replayFact) now applies
preserve-if-absent: a metadata-only after-image replayed over an already-
vectored row carries the existing vector forward instead of deleting it via
writeNounRaw/writeVerbRaw's exact-restore null-means-delete contract (which
must stay exact for transaction-abort rollback). A genuine tombstone still
removes both legs.
A zero-norm vector is lawful inside brainy (cosine distance scores it at
maximum, never a false top hit) but a false attractor for a downstream
engine serving squared-euclidean distance, which cannot tell a real
all-zero vector apart from a legitimate origin point.
- The VFS root now persists with vector [] (the existing "unvectored"
shape) instead of a real all-zero 384-dim placeholder, and is never
routed into the deferred-embed pipeline.
- A one-time migration in the root-init path detects a pre-fix store's
all-zero placeholder root (by norm, not length) and rewrites it to []
through a new sanctioned Brainy method that keeps the canonical
vectored-noun ledger honest and removes the row from the vector index.
- The vector-index write seam (AddToVectorIndexOperation,
ReplaceInVectorIndexOperation, and the generation materializer's direct
insert) now refuses any real all-zero vector before it reaches a
provider, loudly naming the entity, while the canonical write still
lands.
- add()'s dimension-pinning and HNSW-insert gates, and the add-params
validator, now treat any empty vector as carrying no dimension
information, closing a latent trap where an explicit `vector: []`
would have pinned dimensions to 0.
Two builds of the same source tree could publish different artifacts
because buildEmbeddedPatterns.ts and buildTypeEmbeddings.ts stamped
their generated output with new Date().toISOString(). Route both
generators' "Generated:" header and the generatedAt runtime field
through a shared resolver: newest git commit time among the
generator's inputs (script + source data), falling back to the stamp
already present in the previous output when git is unavailable (e.g.
a published tarball build), and finally to a fixed epoch value. Every
fallback logs to stderr so degradation is never silent.
Regenerated both committed output files once so the tree carries
deterministic stamps; no other content changed.
Two cures from the pair's first production adoption, both measured live.
THE READ GATE IS PER-FAMILY. The report-driven gate refused on ANY
provider's not-ready verdict at every read choke point — so a pure
metadata find({ where }) was refused because the VECTOR leg was not
serving, and a deployment's badge reads returned errors for a verdict that
had nothing to do with them. A read may only be refused by the family it
actually consults: metadata reads by the metadata leg (plus graph for a
`connected` filter), vector search by the vector leg, traversal by the
graph leg. Callers name what they need; the existing narration-once-per-
generation and typed-refusal laws are unchanged within a family.
NO RE-EMBED ON UNCHANGED DATA. update() — and its transact() planner —
treated any write that carried `data` as a data change: with
deferEmbedding it queued a landing, and the worker re-embedded and re-landed
a vector for content that had not changed. A host heartbeat re-writing an
unchanged row every few seconds therefore fed a live index-row loop on a
production store. A write carrying the row's current data (structural
compare, key order normalized) is now not a data change: no re-embed, no
deferred landing, no vector rewrite; the metadata write itself still
commits. A real change re-embeds exactly as before.
Pinned in tests/integration/read-gate-scope-and-no-reembed.test.ts — both
pins red-proved against the unfixed code with the production shapes
verbatim. Two health-gate pins that encoded the old brain-global scope are
re-pointed to the family their reads consult.
The coverage denominator the health-by-accounting ratification named for
the vector family — never built until now, and its absence was measured as
the exact outage class it existed to prevent: a migrated store with
canonical vectors and no derived index opened with the vector leg EMPTY,
served [] from vector search with no error, and the report-driven read gate
had nothing to refuse on (the provider's coverage invariant was honestly
unledgered — the denominator was ours to supply).
- getCanonicalCounts() gains vectors: { all } — the count of canonical
nouns holding a REAL vector. Incremented where a vector lands (the
isNew-gated metadata seam for explicit vectors — the same discipline that
keeps HNSW neighbor-link re-saves from inflating counts; a narrow
noteVectorLanded hook for the deferred-embed landing, gated on the
worker's own pre-embed read). Decremented on a proven delete of a
vectored noun; a vector-uncertain delete marks the ledger suspect rather
than guessing (no new reads on the delete path). Recounted by the
sanctioned recount; legacy counts.json derives it once (a deferred noun's
vector file exists with an empty vector, so presence requires one
content read at derivation — never on the hot path).
- The open gate's vector leg: when a health-reporting provider claims
serving while the index holds zero nodes and the ledger proves vectored
canonical rows exist, open BUILDS (narrated) — routed through the
provider's idempotent fillFromCanonical() when exposed (the joint door;
a partial shortfall stays repair()'s operator business), the JS rebuild
otherwise — or fails typed pre-serve. Scoped exactly: bare isReady()
providers, migrating providers, and white-box size stubs open as before.
Pinned end-to-end from the partner gate's probe shape (store with vectored
canonical rows, no derived index, reopen → search serves N, never []),
red-proved against the pre-fix path; the inverse (zero vectored rows) opens
without building and serves [] honestly.
resolveVerbEndpointInts mirrors the resolved u64 endpoint ints onto the
verb object itself as BigInt (verb.sourceInt/targetInt) for the graph legs'
own params. The live verb path's delete legs then reused that same object
as the metadata-index crossing — and the seam's metadata is JSON-safe by
contract (a native provider serializes it; u64 as Number corrupts above
2^53), so JSON.stringify threw and the whole transaction aborted. Found by
the first joint pair gate; four downstream suites red from one crossing.
The crossing now routes through a JSON-safe view that drops BigInt-valued
top-level keys — endpoint ints ride their own op params on the graph legs,
exactly as designed, and never the metadata crossing. Applied at the
retraction helper (cascade + unrelate + transact mirrors) and
updateRelation's remove leg.
Pinned by driving the exact shape (relate resolves ints, remove cascades
the same object) through a provider shim enforcing the JSON contract —
red-proved against the unfixed path (the joint gate's verbatim error),
green with the fix.
repairIndex() acted only on heal:'rebuild' — an invariant asking for the
INCREMENTAL heal (re-post exactly what the ledger names, O(missing), never
a store-sized rebuild) did nothing on brainy's side. A failing 'repair'
verdict now routes to the provider's feature-detected repair(); the
post-heal RE-READ of the report decides success (the acceptance meta-pin's
law — run the named heal once, re-read, nothing may still fail the same
way), and a repair that does not converge is recorded with the escalation
named: repairIndex({ rebuild: [family] }).
The factory loudly rejects every REMOVED pre-8.0 path key, but an unknown
nested `config` object (e.g. `storage: { config: { baseDir } }` — a shape
that was never supported) fell through SILENTLY to the zero-config default
directory. Every instance constructed with such a shape wrote to ONE shared
on-disk root while its caller believed each had its own — found live when
two integration tests' brains shared a store across an entire
single-process CI run and a health probe refused on the foreign edges it
sampled. A nested `config` carrying any path-shaped key now throws the same
loud migration error, naming the canonical `path` rename. The two tests are
repaired to the supported shape (and now actually test isolated stores, for
the first time since 8.0).
validateAddParams() treated '' as falsy and rejected it with "Missing
required field 'data'" — so a legitimate empty file's first write always
failed. Only null/undefined data (with no vector either) is genuinely
absent; '' is real content. Fixed the check, plus the identical bug in
validateUpdateParams() (truncating a file to empty via overwrite hit the
same falsy check) and in update()/transact()'s update planner, where a
plain `Boolean(params.data)`/truthy check on the resolved vector would have
silently skipped both the deferred-embed marker and the eager re-embed for
an emptied value — a stale vector with no path to ever correct itself.
Verified end-to-end: vfs.writeFile('/empty.txt', '') now succeeds,
readFile() returns '', the file lists, and stat() reports size 0; the
existing "should reject empty string as data" tests (unit + integration)
asserted the old buggy behavior and are updated to assert the fixed
contract instead.
vfs.readdir()'s ReaddirOptions.recursive was typed but silently ignored: a
recursive request behaved identically to a non-recursive one. Implemented
properly: recursive listing walks every descendant (files and directories,
any depth) via the same graph-traversal + one-batch-fetch path
getTreeStructure()/getDescendants() already use, and reports each entry as
a path relative to the queried directory (Node's fs.readdir(dir,
{ recursive: true }) convention) — 'sub/file.txt', not just 'file.txt'.
With withFileTypes: true, each VFSDirent.name carries that same relative
path when recursive (matching the string-array form byte for byte);
VFSDirent.path stays the absolute VFS path either way, so no information is
lost. Filter/sort/pagination compose unchanged, now over the full recursive
set. Non-recursive behavior (direct children, named by basename) is
unchanged.
A production restart storm measured 90,017ms for a single brain init vs
1,117ms quiet (~80x contention multiplier), traced to performInit() eagerly
awaiting the process-global WASM embedding engine before the VFS root even
existed. Every writer's open() queued on the one throttled model compile
(90-140s on throttled CPUs).
- VirtualFileSystem.doInitializeRoot() no longer embeds '/'. The root is
system-tier plumbing nothing ever searches; when the default WASM engine
is active it now gets an explicit all-zero placeholder vector
(cosineDistance returns max distance for a zero vector, so it never ranks
ahead of real content). deferEmbedding was considered and rejected: its
landing path kicks the embed worker synchronously right after commit,
which would still force the cold compile within milliseconds — just off
the awaited path, not avoided. A registered native 'embeddings' provider
(no cold-start cost, possibly a different dimension) still embeds the
root for real, via the new Brainy.usesDefaultWasmEmbedder() seam.
- performInit()'s eager-embedding step now only STARTS the WASM engine warm
in the background instead of awaiting it inline. embed()/embeddingManager
already serialize concurrent callers on one shared init promise, so the
first real embed() converges correctly either way; a failed warm narrates
loudly instead of surfacing as a silent latency spike or an unhandled
rejection. eagerEmbeddings: false still means no warm at all.
- FileSystemStorage.init() batches its ~8 independent bootstrap mkdirs
(each creates its own full subtree via recursive:true, so none depend on
the others existing) into one Promise.all. The restore-completion step
and initializeCounts() stay strictly sequential — they have real order
dependencies on rootDir and systemDir respectively.
- performInit() now times five phases (storage init / generation-store
open+fold / index init+gate / VFS bootstrap / embedding-warm-started) and
logs one warning with the per-phase breakdown when total open exceeds
2000ms; silent otherwise.
The auto-detection "not installed" heuristic accepted any resolution failure
whose message merely CONTAINED the package name, unterminated — so a missing
platform-binary sibling package (what a deploy replacing node_modules
mid-restart leaves behind) read as "the accelerator is not installed", and
brainy silently served the default WASM engines with zero journal lines. A
production restart storm paid 90 seconds of throttled WASM compile behind
exactly that hole. The name must now terminate where it ends (quote,
whitespace, punctuation, end) — a sibling package, an inner file path, or a
dependency failure is a broken install and init() throws, as the guard's own
law always stated.
Second door: activate() returning false (the documented graceful decline)
warned on console.warn, which `silent: true` patches away — an invisible
degrade. The decline now narrates via the always-on channel.
Also exports CanonicalCounts from the public surface (the coverage-ledger
denominator type consumers read through getCanonicalCounts()).
Pinned in tests/unit/plugin-activation-loudness.test.ts (five error shapes;
the decline warn under silent: true).
Three cures on the JS metadata index, one seam:
- THE CATCHUP WIRING. The index computed its three-way watermark verdict at
open and nothing consumed it — after a crash + adopt reopen, find() served
the pre-crash index while canonical reads and counts recovered (caught by
the lifecycle lane's first run). The open path now consumes the verdict:
'adopt' is a no-op, 'catchup' folds the fact window (stamped, committed]
through the index legs — nouns and verbs, remove-then-add, one mechanism
for add and update — and 'rescan' runs the explicit rebuild, each narrated.
The lane's Ch4–6 release-blocking marker comes off: the contract holds.
Bonus root-cause: close() never stamped the projection watermarks (only
flush() did), so any close without a prior flush verdicted a needless
'rescan' on reopen — both doors now stamp.
- THE LIVE VERB PATH. Verb rows entered the metadata index only via rebuild
walks, so every rebuilt store minted phantom/stale verb postings from its
first live relate(). relate()/unrelate()/updateRelation() and remove()'s
cascade now post/retract the verb's row in the same commit as the graph
leg — transact() planners mirror identically — using the exact record
shape the rebuild walk uses, so live and rebuilt populations agree.
- THE ONLINE REBUILD. rebuild() was clear-then-walk — every metadata read
empty for the duration. rebuildMetadataIndexOnline builds a fresh manager
beside the serving one (shared identity, in-memory build, dual-write via
a shadow seam with zero call-site changes), atomically swaps the
reference, and persists exactly once post-swap. A find() polled ~200x
during a 2k-noun rebuild never dropped below its baseline.
repairIndex({ rebuild: ['metadata'] }) uses it automatically.
The read gate stops consulting the unnamed isReady() boolean: every provider
may expose healthReport() (sync, O(1), composed from exact ledgers —
HealthReport with a monotonic generation, per-invariant source
ledger|deep|unledgered, missing {count, sample}), and one readiness
authority (assessProviderHealth) derives the verdict. Unledgered families
are UNKNOWN — never healthy, never broken; a report that throws is a loud
not-ready, never a shrug. Reads at the four index choke points refuse with
the typed NotReady errors, narrated once per (provider, generation) — a
read NEVER starts a store walk:
- the first-read lazy build retires (open builds instead, regardless of
size — the ≥10k deferral and the "lazy loading on first query" branch go;
disableAutoRebuild is re-meant honestly in its docs);
- the verify*Live read-path rebuild triggers retire (refuse-or-serve);
- the read-time consistency probe that could launch a dark rebuild from an
ordinary find() retires;
- repairIndex({ rebuild: ['metadata'|'graph'|'vector'] | 'all' }) is the
one explicit door: rebuilds the named leg unconditionally and reports
rebuilt per family; bare repairIndex() stays report-driven.
test(lifecycle): the biography lane — a store's whole life, refereed
tests/lifecycle/: an independent shadow model referees every read after
every chapter (founding, a working day, clean restart, crash, repair,
second life). Chapters 1-3 green. Chapters 4-6 assert the true contract and
are marked .fails as a release-blocking finding (the kill-matrix
convention): after a crash + adopt reopen the metadata index computes its
'catchup' watermark verdict and nothing consumes it — find() serves the
pre-crash index while canonical and counts recover. The catchup wiring is
the cure; a passing .fails will force the marker's removal. The lane runs
in the integration gate (config + coverage guard).
A prerelease the other engine devDeps from our own registry while the two
are being proven together must never reach npmjs: --source-only tags, lets
CI publish to The Source, creates the release page, and SKIPS the storefront
publish, the cross-registry verification and the docs push — loudly, at
each step. Refused for a non-prerelease version: a public floor always ships
byte-identical on both registries.
The pin wrote a post-flip row, abandoned the brain as crashed, and then
deleted the row's canonical bytes to prove the first post-flip boot folds
BOUNDED above the flip's stamp. Between the write's ack and the abandon sat
the store's 50ms pending-flush timer: on a loaded box (the plant lane) the
flush won, barrier-synced the row and advanced the checkpoint over it — and
the fold, correctly bounded, did not restore bytes the test had destroyed
after they were stamped durable. Green locally, red on the plant: the
engine was right, the pin was timing-dependent.
The crash is now armed at exactly singleop-after-fact-append: the fact is
appended and at-ack synced, no flush is ever scheduled, the stamp provably
still reads the flip's value when the pre-flip bytes are dropped, and the
post-flip row's bytes — which lived only in the pending tier's RAM — are
lost for real, not synthetically. The reopen must re-materialize it from
its fact and must not restore the pre-flip row.
A validateInvariants() that threw was re-synthesized by the host's catch as
heal:'rebuild' with serving:false — a rebuild lever one transient exception
away, while the native provider's own composer reports the same event as
heal:'none'. Two components disagreeing on what a thrown check means is how
a flaky probe becomes an outage. Both now agree: the report is named
('validate-invariants-threw'), loud (healthy:false, the error in detail),
unverified — and it never buys a rebuild and never withholds serving; the
provider's serving verdict is the provider's to compose, not inferred from a
probe that failed to run. The synthesized report also carries the provider's
name instead of 'unknown'.
RepairFamilyReport gains `missing: {count, sample}` (an exact count plus a
capped id sample — a verdict, not a dump), `rebuilt` (a full generational
rebuild ran, as opposed to an incremental heal) and `reason`, aligning the
receipt's shape with the provider-side health report.
Pinned in tests/unit/validate-invariants-delegation.test.ts.
The storage-level unfiltered getNouns()/getVerbs() walks enumerate every
tier, but their totalCount reported the user-facing scalar, which skips
system/internal records on the write path — so a derived-index coverage
ledger comparing its posted count against that total would read
"over-posted by N" on every store with a VFS. This adds the ledger's real
denominators:
- totalNounCountAll / totalVerbCountAll: +1 for every new canonical record
regardless of tier, −1 for every PROVEN delete (record read, or the
caller's prior image), persisted in counts.json beside the counted
scalars, recomputed by the sanctioned recount (rebuildTypeCounts).
- The unfiltered storage-level totalCount is now the ALL scalar and is
never clamped: Math.max(scalar, scanned) could only move a scalar up, so
an inflated counter hid forever; a divergence is now visible and healed
by repairIndex().
- A delete that cannot prove the record existed never decrements on faith:
it marks the ledger SUSPECT (persisted, narrated once per session) and
the recount clears the flag with proof.
- getCanonicalCounts() on StorageAdapter (optional) exposes {counted, all}
per family plus the suspect flag — O(1), no I/O.
- A counts.json written before the ledger existed derives both scalars
once from the canonical id tree at open and persists them; absent keys
are a legacy file, never a zero.
User-facing getNounCount()/getVerbCount() are unchanged.
Pinned in tests/integration/canonical-count-ledger.test.ts (5 laws).
remove() guarded its index legs with 'if (metadata)': a row whose canonical
metadata was unreadable at delete time kept its postings forever, silently
(the leak class a partner audit confirmed at this exact site). Closed at
all three provider shapes: a provider exposing the id-keyed removal
contract (removeEntityById, arriving with the accelerator's next minor)
gets exact per-entity retraction; the JS index gets id-keyed cleanup
(deleted bitmap + id mapper — field stats reconcile at rebuild), narrated;
a native provider without the contract is NEVER called metadata-omitted
(that path walks the store's value space) — its skip is narrated and
tracked in the degraded set for repairIndex, never silent. The pre-reads
are torn-tolerant: a torn record is deletable (the delete is the cure).
Pinned: a metadata-less row with live postings deletes cleanly and leaves
the query universe; the delete-family suites stay green alongside.
repairIndex() now returns a RepairReport: one row per repair family
(orphaned containers, count rollups, VFS containment, metadata corruption,
write quarantine, each provider's invariant pass, degraded-read state) with
checked / healed counts and an explicit skip reason for anything not run —
no silent rows. A summary line narrates families checked and heals applied.
This is the receipts half of the graph-trust program's ask: a repair that
cannot show its work per store is a repair nobody can audit. Pinned: a
healthy store yields a complete zero-heal receipt with every family
accounted; a manufactured pre-8.3.1 ghost container appears in the receipt
as a counted heal. Additive: void-callers are unaffected.
A production store acked writes while readback served empty for fifteen
minutes. The brainy half: the lazy readiness gate had exactly one caller —
find() — while related() and every VFS path served straight from providers
that init had deferred (disableAutoRebuild on a large store). A false
health verdict from the accelerator pulled the trigger; the unguarded read
surfaces were the gun.
Every index read funnels through three helpers; the gate now lives at those
choke points, so any first read on a cold instance waits for the build and
then serves truth — the fast path after the latch is one boolean. The lazy
build's start is narrated through the production logger, never the
silent-suppressible console: fifteen silent minutes taught that line.
Pinned with the production shape: related() as the first-ever read on a
fresh lazy instance serves the relation; a filtered find serves the row.
Two guards for every gate lane, born from the 2026-08-13 lost-day ledger.
gate-preflight.sh refuses a lane on a machine that cannot be trusted to
produce honest numbers — co-tenant processes named by pid and command, load
average, CPU governor, disk floors — one FATAL line per violation so the
operator can act from the message alone. vitest-verdict-check.sh refuses a
suite log that cannot be trusted as a verdict — missing or mismatched
summary counts, files that never executed (a truncated run once read as
green from three files of ninety-nine), and worker-pool death signatures.
Both verified live: the preflight correctly refuses this workstation naming
its actual offenders; the verdict guard passes/fails five fixture shapes
(clean, wrong-count, truncated, worker-death, no-summary) and both CLI
modes. Wire-up into the CI lanes rides the runner program.
A production brain's first process boot after a live authority flip looked
hung and was restarted three times mid-recovery — three defects with one
scene. (1) THE FOLD MATERIALIZED THE LOG: peekFactsAbove(0) decoded every
fact into one array (GBs of after-images on a ~7k-fact log, a GC storm, a
starved write lane). The fold now STREAMS one segment-batch at a time —
memory is one segment at any log size — with structural ordering asserted
loudly. (2) THE FOLD WAS SILENT UNTIL DONE: minutes of boot work with zero
narration is what invited the restarts. It now announces itself BEFORE the
work ('do not restart, the fold is finite') and prints progress every
thousand facts. (3) THE CHAIN COULD ONLY ARM AT A CRASH: a live mid-session
flip left the fold checkpoint unfounded, so the brain's first unclean boot
paid a whole-log fold. Adoption now founds the checkpoint AT THE FLIP — one
paged full canonical barrier (bounded memory), then the stamp — so bounded
recovery holds from minute zero for every store that flips, at any size.
Pinned: a non-fresh flip stamps immediately; the first post-flip unclean
boot folds bounded (an unflushed at-ack fact above the checkpoint is
restored; a barrier-covered row below it is outside the fold). Kill matrix
and both adoption suites green alongside.
The plant's integration lane caught it twice: the fence's startedAt-strict
comparison turned the documented same-process warn-and-take-over path (two
instances in one Node process — the server-restart test pattern, and the
shared-default-store pattern across test files) into a flush-killer: the
first instance's background flushes latched dead while its own process held
the lock ('PID N no longer holds the lock — it is now held by PID N').
Ownership is per-process: pid + hostname. startedAt stays in the lock for
observability but not in the fence — it protects nothing (a pid-recycled
successor's victim is a dead process that runs no fence checks) and it
convicted the innocent. Pinned: a same-process re-open leaves both
instances' flushes working; the cross-process eviction pins unchanged.
Verified under the lane's exact command: 102/102 files, 850 passed, exit 0.
Seven micro-budget tests were calibrated on one fast desktop and failed on
other honest iron with zero functional failures (bisect-proven pre-existing;
David-waived for 10.1/10.2 with this recalibration filed as the cure). Every
budget is now at least 3x the worst measurement observed across three
machines, each with a comment naming its calibration basis; the find-unified
micro-comparison of two sub-millisecond timings becomes a ratio assertion
(absolute equality of microsecond pairs can never be stable). The
inference-bound trim-history correctness test gets a timeout covering its
slowest observed run (174s) — its assertions are exact and untouched.
These remain order-of-magnitude guards; real perf enforcement lives in the
dedicated perf lanes with iron-specific budgets, per the gate-speed standard.
Known non-test artifact, documented not hidden: on slow-inference machines a
minutes-long awaited-embed loop can trip vitest's worker-RPC 60s tolerance
('Timeout calling onTaskUpdate') — all tests pass, vitest exits 1 on the
unhandled orchestration error. The CI lanes on faster iron exit clean; if a
lane ever trips it, the test moves to deterministic embeddings (its
assertions are size-bookkeeping, not embedding quality).
The production dev-store split-brain (two live writers alternating a store's
id-mapper between two internally-consistent truths), cured at all three of
its roots. (1) STALENESS REQUIRES PID-DEATH: the old rule evicted on
heartbeat age alone, so a >60s event-loop stall (debugger pause, GC, heavy
sync work) handed the lock to a second opener while the first kept writing;
a live process is now never auto-evicted — a wedged-but-alive holder is the
operator's call via {force:true}, and the heartbeat stays for observability.
(2) THE CLAIM IS ATOMIC: writeFile(wx)'s open→write→close left an empty-file
window a concurrent opener could read as torn, unlink a LIVE claim, and take
the lock; the claim is now tmp-write + hard-link — the lock appears with its
full contents in one step. (3) THE FENCE: every flush commit and transact
barrier verifies lock ownership first (one small read per window) — a
forced-out or lock-deleted writer fails typed (BRAINY_WRITER_FENCED) before
a single staged byte or manifest advance, instead of writing on unaware.
Pinned: live-with-ancient-heartbeat refuses typed; dead-PID self-clears
narrated; a forced-out writer's flush and transact both fence, advancing
nothing. Requested by a downstream team as single-writer guard or loud
lockout — this is both.
Two consumer-driven cures sharing one stamp. (1) TX-LOG ORIGIN: engine-
originated commits stamp an optional origin on their tx-log entry AND the
commit fact's meta — 'system:embed-landing' (the deferred vector landing),
'system:adoption-backfill' (baseline re-commits), 'system:reconcile'. A
downstream activity feed showed a double tick because the landing commit was
indistinguishable from a user save, and the consumer rightly refused a
time-window collapse as a quiet loss; feeds now filter on fact. User writes
stay unstamped — absent origin is the user shape, every existing consumer
unchanged. (2) reconcileLogDivergence(id, {attest}): the human's door for
log-live-canonical-absent, the one class adoption refuses by design because
a lost-tombstone deletion is indistinguishable from canonical loss.
'deleted' mints the missing tombstone (history keeps the earlier live
record); 'restore' folds the log's only copy back into canonical; wrong-
class calls refuse typed with nothing written. Loud, narrated, single-row,
origin-stamped. From a production adoption's one surviving divergence.
A production brain with a 12.7k-row pre-log baseline advanced exactly 800
rows per adoptLogAuthority() call (a five-pass ceiling × the oracle's
200-row listing cap), refused the flip, and sat tree-authoritative for
hours across restarts. The bound was sized for drift, never for a baseline.
Now: the adoption path runs the oracle uncapped so ONE scan yields the
ENTIRE curable set, every pass cures all of it, and the loop runs to
completion with the no-progress guard as its only stop. Pace rides the
write path (one full-brain scan amortizes over thousands of cures, not two
hundred): 1,000 drifted rows adopt green in one call in ~10s. Progress is
narrated for a live operator. The wire report keeps its 200-row cap.
Pinned: a baseline above the old ceiling adopts green in a single call.
Two defects with one root, found by the fold-checkpoint work's first
integration gate. (1) THE RACE: restore() never quiesced the generation
store, so a background flush could write into _system/ while the swap was
removing it — observed as ENOTEMPTY mid-swap when a checkpoint stamp landed
between readdir and rmdir. The swap now runs inside the store's exclusive
section (runStateReplacement): flush timer disarmed, pending tier and
checkpoint accumulator discarded BEFORE any directory moves. (2) THE
INHERITED ASSERTION: a snapshot carries its source brain's clean-shutdown
marker and fold checkpoint, but the restored files were bulk-copied without
per-file fsync — the inherited stamps would suppress exactly the recovery
fold that cures a post-restore power cut. reopenAfterRestore now deletes
both stamps before reopening: the open treats the store as uncleanly shut,
folds the restored log into canonical, barrier-syncs what it re-applied,
and stamps fresh — the restored state is durably founded at restore time
instead of borrowing assertions about bytes this disk never synced.
Pinned: restore under in-flight traffic completes; the pre-restore stamp
does not survive; the post-restore stamp is the reopen fold's own, at the
restored watermark.
The fold checkpoint (_system/fold-checkpoint.json) is stamped strictly after
a canonical-sync barrier over every live entity touched since the last stamp
(syncEntityCanonical: ids → canonical paths → fsync; an absent file fsyncs
its parent directory so deletes are as durable as writes). An unclean open
under log authority now folds only (checkpoint, head]; the chain bootstraps
at an empty brain's adoption (three-phase hooks around adoptLogAuthority) or
at a brain's first whole-log fold — existing brains converge at their first
crash with zero regression. Rollback restores sync immediately; abort paths
feed the barrier; a failed barrier retains the old bound (bigger fold later,
never a lost write). Five structural pins including boundedness itself.
Also: the production-shaped write-flow gate leg (mixed traffic racing
flushes, crash mid-traffic, every ack survives — from a consumer-reported
gate miss), and two release-ceremony cures (tag-first push so the publish
never queues behind the release commit's CI run; raw-curl npmjs shasum
probe with propagation grace instead of a one-shot false divergence).
An adopter's full suite found two v2 write-path defects on fresh brains,
reproduced with stacks; both cured and both pinned with their exact
production shapes:
1. PAD-FRAME CONSTRUCTIBILITY: a single msgpack bin filler steps its
header by one byte at each size class (bin8→bin16→bin32), leaving one
unreachable payload size per boundary — the sealer requested a
291-byte pad, the encoder threw 'not constructible', and sync() died
whole. Construction is now TOTAL: the class-boundary holes bridge with
a trailing fixint beside the bin ({bin(n)} ∪ {bin(n)+fixint} covers
every size ≥ minimum). Pinned exhaustively: every size from the
minimum through a full sector plus boundary spill constructs
byte-exact and decodes as reader-invisible filler.
2. THE NON-MONOTONIC REFUSAL LOOP: the append-failure compensation
rewound the generation counter on ANY throw — including a covering
SYNC failure after a SUCCESSFUL append. The log carried generation N
while the counter re-minted N, and every later append refused
'non-monotonic (N ≤ head N)' — the write path wedged in a refusal
loop through deferred-embed retries and flush backoff. The
compensation now splits by phase: an append failure (log never took
the fact) fully compensates — un-buffer and rewind; a sync failure
after append earns the rewind ONLY if the appended fact is provably
dropped, otherwise the generation stays consumed and buffered — the
counter never re-mints a number the log may carry. Pinned: an
injected one-shot sync failure fails its write loudly and the very
next write mints fresh and succeeds, with the log scanning strictly
ascending end to end.
Also probed against the adopter's carried report: the 9.0 vfs.rename
stale-ghost shape does NOT reproduce on this head (old path cleanly
unresolvable on exists/stat/readdir after rename).
Gates: unit 2067/2067 (160 files) · integration 833 (97 files) ·
conformance 36/36.
A first adopter's namespace migration went 341 red on one class: the
never-carried-field refusal firing on CORRECT filters against fresh and
sparse stores — a freshly provisioned tenant refused its own first
filtered read, with the did-you-mean built for typos firing hardest on
day-one stores where nothing is wrong.
The ruled cut: a WHERE filter naming a field no row carries is SERVED
OPERATOR-TRUTHFULLY — eq/in/range/contains answer [] (nothing carries
it, nothing matches); ne and exists:false answer ALL rows (the equally
true complement — a blanket empty here would be silently wrong, which is
why the simpler cut was rejected); exists:true answers []. Served from
the field registry, with the did-you-mean demoted to a once-per-field
WARN. orderBy and genuinely ambiguous addresses KEEP their hard typed
refusals: no truthful order exists over an uncarried field, and
ambiguity is a contract error while absence is data.
Mechanics: the negative operator absorbs the FIELD_NOT_INDEXED throw as
its empty exclude set (the clause-level catch correctly zeroes positive
operators only); the egress matcher already agreed. Plus the
provider-seam belt: a field refusal thrown by a replacement metadata
manager is normalized to THIS package's UnresolvableFieldError at every
filter call site — one class identity for consumers, instanceof works
(a first adopter's cross-package finding).
Conformance: tests/conformance/sparse-store-cut.test.ts — the shared
operator rows both engines run (positive-empty, negative-all,
fresh-tenant day-one, orderBy refusal kept, compound composition).
Gates: unit 2065/2065 · integration 832 · conformance 36/36.
The last rung of the default-flip ruling: with the sentinel exemption in,
real production-shaped brains still refused adoption over state-differs
mismatches the backfill could not cure — rows written before the
hydration law carry denormalized wrapper fields that disagree with their
own metadata leg, and the previous as-is identity re-commit PRESERVED
that drift, so the oracle re-flagged it every pass and the flip never
happened. In practice the crash-safe default reached zero existing
brains: the exact outcome the hold ruling forbade.
The cure: the backfill now rewrites canonical in the LAW SHAPE — exactly
the wrapper the log's reconstruction produces (denormalized enumeration
fields derived from the metadata leg, which is their authority under the
field-addressing law; the embedding floats ride through byte-identical;
adjacency residue keeps its own rebuild path). The oracle then verifies
the rewrite before the flip — the same safety, no operator chore.
Log-ahead divergence classes (a log the witness denies) still refuse
loudly, exactly as before.
Classification note for the record: the flagged uuid-v7 rows postdate the
fact log's introduction, so they classify as state-differs (in-log,
drift-shaped) rather than pre-log — both classes ride the same backfill.
Pins: a manufactured depot-shape drifted wrapper adopts green with floats
preserved and metadata intact; log-ahead still refuses typed.
Gates: unit 2065/2065 · integration 832 · conformance 31/31.
The release-holding finding from the joint gate's six real depot brains:
the adoption path's positive-int mint check false-flagged the reserved
VFS-root sentinel (the all-zeros UUID, minted int 0 BY CONSTRUCTION at
genesis on existing brains) as a corrupt mint — so every existing brain
refused log-authority adoption and stayed on the old lossy-under-power-cut
durability, defeating the release's headline crash-safety exactly where
it matters most.
The exemption, at both mint seams (the host's minter thunk and the fact
log's encoder guard): int 0 is legal iff the id is the reserved root;
zero for ANY other id remains a corrupt-mint refusal naming the reserved
exception. The codec's u64 layer already tolerated 0 — only the guards
over-refused.
Pins: adoption goes green on a brain whose VFS root carries int 0 (the
depot-brain shape, previously refused) · a non-root zero still refuses
typed at the mint seam — held at the seam itself because a full write
SELF-HEALS a poisoned zero (the index cycle re-mints before the fact is
written, which is the correct outcome and was verified in the pinning).
Gates: unit 2065/2065 · integration 830 · conformance 31/31.
The quiet-loss cure regressed recovery: the new typed torn-record error
was correct at identity-read time but threw inside init-time recovery
walks, killing opens that previously survived. The boundary, redrawn:
- IDENTITY READS (get-by-id of a specific record, CAS blob point-get):
typed TornRecordError, unchanged — a caller who asked for THAT record
can act on the answer.
- SET-SHAPED READS AND WALKS (enumeration, pagination, batch hydration —
the paths recovery rebuilds and finds page over): HEAL PAST the torn
victim. The adapter's loud floor (error log + counted gauge) fires at
the encounter; the walk serves the remaining rows. One crash casualty
can no longer kill every query on its shard — or the open itself.
- WRITES OVER TORN RECORDS ARE THE CURE: the save path's read-merge, the
commit path's before-image capture, and the operations' rollback
captures all treat a torn prior as the create sentinel, narrated — the
incoming bytes replace the unreadable ones, and history for the id
honestly restarts at that generation. Corruption can never block its
own heal.
- THE NaN SOURCE: torn mapper state (nextId/entries carrying garbage)
discards with narration and re-derives via the existing rebuild path;
the mint gains a source guard healing a non-integer counter from the
live map. The reopen and first-write RangeError shapes are dead at the
source, both authority branches.
Pinned with the exact fault-injection scenarios: a torn entity record
(including the VFS root) no longer kills the open — walks heal past it,
the keeper rows serve, and the identity read of the victim itself is
typed-or-healed; a torn mapper reopens and mints sanely on the first
post-recovery write.
Gates: tsc 0 · unit 2065/2065 · integration 828 · conformance 31/31.
THE DEFAULT FLIP (ruled on proven evidence — at-ack survived 301/301
acked-writes-through-power-cut in block-layer fault injection; deferred
tree authority demonstrably loses flush-covered acks): a brain with NO
stored authority artifact now ADOPTS LOG AUTHORITY AT OPEN. The oracle
gates the flip exactly as the guarded adoption path always did — curable
divergences baseline-backfilled, the flip lands ONLY on a green verdict —
and a brain that cannot verify STAYS tree-authoritative loudly, with the
refusal recorded on the switch artifact so subsequent opens are cheap.
config logAuthority: 'defer' is the explicit documented opt-out (no
automatic adoption; declared flush-window loss; adoptLogAuthority() flips
later). A stored artifact always wins. RELEASES.md carries the posture.
Two standing .fails debt pins FLIP TO HOLDING under the default: the
at-ack crash-survival gap and the ack-at-log durability target — both now
permanent asserted truths, not aspirations.
POWER-CUT THROW SITES (fault-injection findings, brainy-alone config):
- A manifest-listed-but-unloadable column segment QUARANTINES at
discovery (loud once, counted always, quarantinedSegments() exposed for
the heal) and the field serves its remaining segments DEGRADED — never
a raw throw killing every query on the field. Real storage faults still
propagate untouched.
- Torn generation artifacts (NaN/garbage in manifest or counter) DISCARD
with narration at the store's open and recovery re-derives — plus a
defensive finite-integer guard at the init consumer. Never a RangeError
killing an open.
THE LOUD TORN-RECORD CONTRACT: an existing-but-unparseable stored record
now surfaces as a typed, counted TornRecordError on every entity-read
surface (including fifteen previously-blind per-item batch catches);
ENOENT stays clean-absent; artifact readers with designed absent-recovery
keep null-tolerance behind the loud floor. Disk corruption can no longer
read as silent data invisibility.
Suite migration: the default's pins inverted deliberately, generation
baselines made relative, quarantine-contract pins rewritten to the ruled
behavior.
Gates: tsc 0 · unit 2065/2065 (159 files) · integration 826 (93 files) ·
conformance 31/31 · kill-matrix 15/15 · torn-open guards 2/2.
An internal cross-engine fault-injection run (frozen-platter power-loss
capture) surfaced three release-gating findings; each cured in its owning
layer, each pinned:
1. WHOLE-LOG REPLAY ON UNCLEAN OPEN (the big one): log-authority replay
only covered facts ABOVE the manifest — but live canonical entity
writes are tmp+rename without per-file fsync, and the group-commit
flush syncs staging + manifest, never the live tree. Power loss could
therefore vaporize acked canonical bytes BELOW the manifest while the
log held every fact scan-clean (measured: 299 of 301 acks lost).
Now: a clean close stamps a clean-shutdown marker (fsynced, written
last); every open consumes it; an UNCLEAN open under log authority
folds the ENTIRE log into canonical — whole-entity after-images make
the re-apply idempotent and byte-safe. Zero cost on the happy path;
crash recovery pays one narrated fold. Recovery is replay: a crash is
just bigger lag.
2. TORN WRITER LOCK: power loss legally leaves the lock file present but
empty; the parse failure read as 'no holder' while the O_EXCL claim
EEXISTed forever — a PERMANENT lockout no staleness check could clear.
An unparseable lock is stale by definition (no live holder has one):
unlink loudly and re-loop; a racer rewriting a valid lock first wins.
3. PAIR GUARD: flush() called metadataIndex.stampWatermark unguarded;
a replacement metadata provider without the method killed the pair at
first flush. All three stamp calls are optional-chained — a missing
stamp is a verdict-side rescan, never a flush crash.
Pins: whole-log fold restores rows vanished below the manifest ·
clean-shutdown marker lifecycle (stamp/consume/re-stamp) · torn-lock
recovery with a fresh write after · stampless-provider flush.
Gates: unit 2055/2055 · integration 824 · kill-matrix 15/15.
An internal report from cross-engine write-path instrumentation: with
individual writes slower than the idle window (a contended disk), every
inter-write gap looked idle and fired a background full flush — 15 extra
flushes during 100 contended adds, amplifying the very pressure that
slowed the writes. The law now: an idle fire landing within the spacing
floor of the last flush DEFERS to the floor boundary instead of flushing;
the floor is min(interval, 10× the CONFIGURED idle window) — scaled to
caller intent (a tiny idle window keeps fast idle-driven durability;
default 2s/30s config gets a 20s floor), derived from the configured
idle, never from a deferred re-arm delay (which would compound into
runaway deferral). Deferred is never dropped: a lone write on a
then-quiet store still persists at the floor without any further write
arriving.
Pins: the contended-shape pin (six slow-spaced writes fire ≤2 idle
flushes, not one per gap; then still persist) + the original quiet-store
idle pin unchanged. Unit 2055/2055.
The generic reprojection engine (pure TS; the twin of the native
implementation — same frozen contract, one shared conformance intent):
register any ProjectionAdapter; advance(family, {budgetMs}) folds facts
from the adapter's own watermark to the head in installments ≤50ms with
real macrotask yields; foreground door traffic bumps the DoorSignal and
an in-flight advance yields within one installment ('preempted');
advanceAll round-robins families fairly. swap(family, buildAdapter) is
the doors-open migration primitive: the OLD projection keeps serving
while the new one builds beside it, the flip is atomic at parity, and a
concurrent second swap refuses typed. A fact the fold cannot apply
(typed ProjectionApplyError) is QUARANTINED — skipped, ledgered,
narrated per-doubling, exposed for refuse-affected-reads — the service
class law's fourth answer: never a wedged rebuild, never a silent skip.
The engine never writes stamps: each adapter owns its durability and its
stamp-after-data discipline. Upgrade, heal, and rebuild are now the same
machinery behind open doors.
FactLogSource wires any host's fact scan in one line
(factSourceFromHost(brain)); window-contract violations are loud.
Pins: 23 unit (budget resume without refold · preemption within one
installment · round-robin fairness under a skewed backlog · build-beside
visibility mid-swap · atomic flip · single-flight refusal · quarantine
skip/ledger/doubling · non-typed throw aborts · losing adapter
discarded) + 3 integration on a real brain (fold matches ground truth ·
doors answer mid-fold with the preemption path exercised · crash
mid-fold resumes from the stamp, never refolds).
Gates: unit 2054/2054 (157 files) · integration 820 (93 files) ·
conformance 31/31.
The private recovery discipline, applied to its own machinery: pending-
embed markers stop being sidecar files and become first-class log records
riding the write's OWN commit fact — embed.pending lands in the same
atomic append as its after-image (a marker can never be orphaned from its
write, or vice versa; in durable-at-ack mode it shares the write's
covering fsync — zero extra syncs), and the worker's landing commit rides
embed.landed with the inline vector. Crash recovery is now a FOLD of the
log (pending without a matching landed = recovered), skipped wholesale on
brains with no v2 history; the one-time legacy bridge folds existing
sidecar files in, migrates them as one fact, and deletes them —
idempotent under a crash mid-bridge. No code path writes the sidecar
again.
Plus the ENTITY-TRUTH digest law, found by this train's own pins:
canonical vector wrappers denormalize HNSW residue (connections + the
randomly-assigned node level) that the log deliberately does not carry —
the verification oracle digested it and would have reported false
state-differs on ~any nonzero-level node (a ~15% flake in the cutover pin
was the symptom). Both sides of every oracle comparison now normalize to
entity truth (nounEntityTruth); index residue has its own rebuild path
and is not entity state.
Pins: embed-markers-in-log 5/5 (same-generation marker, landed+fold-to-
zero, crash recovery via the log with the sidecar prefix EMPTY on disk,
legacy bridge, VFS hung-embedder ack) · deferred-embedding 5/5 unchanged
(the contract outlived its mechanism) · kill-matrix 11/11 · cutover 5/5
×10 runs (flake dead) · unit 2031/2031.
One deterministic v2 log (nine facts covering every fold-relevant behavior:
genesis, after-images with minted ints, a deferred embed pending→landed,
a sameAsGeneration vector ref, a verb, a tombstone, and an all-deduped
empty commit) whose ENCODED BYTES and FOLDED STATE are both pinned by
sha256 literals. The fixture (tests/fixtures/golden-log-v2.bin, 4128 B,
byte-verified against the encoder on every run) is the shared artifact a
second reader implementation consumes — it must reproduce the identical
fold digest; the pair is normative on disagreement. The fold law is
stated in prose beside the code: generation-ordered latest-per-id,
tombstone masking, embed.landed vector application, single-hop ref
resolution, key-sorted digest.
Also: decodeGroupV2 discriminated pad filler by RECORD COUNT, silently
swallowing legitimate empty commits (an all-deduped batch at a real
generation). Pads carry generation 0 — which writers can never mint — so
the generation is the honest discriminator; empty commits stay visible.
Pins: 4/4 (encode-exact, fixture-identical, fold-exact, human-readable
spot checks beside the hashes).
- Watermark stamping fans out at flush: all three projections stamped
with the committed generation before their flushes persist.
- waitForIndexed(path?, {generation, timeoutMs}) — the one honest read
barrier for write-then-recall consumers; typed timeout error carries
the pending count and names the gauge; getIndexStatus() gains
per-projection gauges. awaitPendingEmbeds() unchanged underneath.
- adoptLogAuthority() self-backfills curable divergences (pre-log
records, witness drift) by identity re-commit before flipping — a
fresh brain flips clean; log-ahead divergences still refuse loudly.
- The verification oracle gains VERB legs (all four divergence classes;
unwired = honest verbsChecked: 0, never a scope claim).
- find({where: {}}) match-all serves (was silent-empty, warm AND cold;
same fix in count/streaming/subgraph seeding); removeMany({where:{}})
refuses typed — a match-all bulk delete must be explicit.
- Aggregation native envelope stamped via noteSourceGeneration before
serializeState; the native-blob restore gates through the same
adoption verdict as caller-side state (the unconditional adopt dies).
- LC8 pinned: a wholesale directory move opens and serves identically
across all three intelligences, with history traveling.
Gates: unit 2031/2031 (156 files) · integration 812 (91 files) ·
conformance 27/27.
Every persisted projection artifact (metadata field indexes + column
segments, HNSW node records, graph adjacency LSM trees) now carries a
stamp asserting 'this state reflects every committed generation ≤ W,
atomically' — written LAST in each owner's flush (stamp-after-data: a
crash between data and stamp = unstamped = rescan, never trust). At load,
each owner computes the three-way verdict: stamped==committed → adopt
(zero work) · behind → catchup (gap reported) · above/unstamped → RESCAN,
loudly. Legacy artifacts re-derive once, then are stamped forever. Shared
law in projectionWatermark.ts (the aggregation verdict machinery,
generalized); vector artifacts carry model dimensions. Verdicts are
computed and exposed (watermark()/watermarkVerdict()/watermarkGap());
rebuild triggers unchanged — acting on 'catchup' is the fold train.
Pins: 22 unit (7 metadata · 8 hnsw · 7 graph, incl. spy-order
stamp-after-data) + the end-to-end reopen-adopts pin.
The cutover: new tail segments write format v2 (per-record [type, version,
cipherFlag, keyId] envelope; noun/verb after-images carry dense ints
MINTED AT APPEND from the id mapper — a rebuilt mapper reproduces
assignments exactly; log.genesis opens every new log with the id-space
width + a minted brain id; sync() seals to the header-declared sector
boundary with reader-invisible pad frames). Existing v1 segments are
never rewritten — per-segment decoder dispatch reads both formats and v2
facts map to the exact CommitFact shape all consumers already read.
Cutover on a live v1 log: an empty v1 tail re-heads in place; a non-empty
one is sealed by rotation, byte-identical. Records reserve the encryption
fields (cipherFlag 0 / keyId nil are the only legal values; anything else
refuses typed naming the needed newer reader) — crypto-ready with no
future bump on the compat surface. Empty-records facts are legal (an
all-deduped batch is a real generation — v1 semantics preserved; the
refusal there tore a column-store flush mid-commit in the full suite, the
consistency guard caught it loudly, and the root is fixed).
Golden byte vectors pinned for the second (native) reader implementation.
Pins: cutover 5/5 · codec 54 · kill-matrix stays 11/11.
The time-travel recall row moves from envelope-note to contracted: vector
search at a pinned past generation serves the vectors AS THEY STOOD —
a later re-embed never leaks into an earlier pin (byte-exact), tombstones
mask, the deferred-embed pin serves the stub on the vector leg until the
landing generation (text/metadata legs unaffected — triple intelligence by
design), and beyond-head pins refuse typed. Brainy-alone leg = the
documented ephemeral at-generation materialization; the at-scale leg rides
the accelerated provider's as-of index. Registry row added (shared ID
pending the master table).
Two release-blocking findings from the durability kill-matrix, both fixed
in the owning layer:
1. LOG-AUTHORITY REPLAY AT OPEN: durable-at-ack fsynced the fact before
the ack, but open() truncated every fact above the manifest — after a
power loss that takes the un-fsynced tmp+rename canonical bytes, the
acked write's ONLY durable copy was discarded. Now: under 'log'
authority, open() REPLAYS intact facts above the manifest into
canonical (FactLog.peekFactsAbove — CRC-gated, order-sorted) and
advances the manifest to cover them; tree-authority brains keep the
truncate contract they were promised. Pinned end to end: the power-loss
row constructs the exact disk state (fsynced log, vanished canonical
rename) and the acked write lives.
2. NO SILENT COMMIT: commitSingleOp buffered the generation BEFORE the
fact append; an append failure (ENOSPC) rejected the caller but the
next flush durably committed the generation with NO fact — a permanent
silent log gap. Now the failure path un-buffers and returns the counter
reservation: nothing commits, the log stays gap-free, and the canonical
execute-residue orphan is the documented crash-equivalent.
Plus: the kill-matrix itself (11 rows — every commit-path fault point ×
reopen-as-crash recovery contract, at-ack variants, disk-full row; five
new zero-cost faultPoint sites), the log-authority pin suite (oracle
green/red/state-differs, flip refusal, switch survives reopen, 9/9), and
the group-commit covering pins (5/5).
Gates: unit 2002/2002 (152 files) · integration 785 · conformance 27/27.
The provider contract (metadata addToIndex/removeFromIndex, vector
addItem/removeItem, id-mapper getOrAssign/remove) gains an optional
trailing generation — evaluated lazily at operation execute time (the
graph surface's thunk pattern, generalized), threaded from all 17
construction sites: undefined during generation-0 bootstrap, the real
committed generation everywhere else. Optional = additive: no existing
provider or caller breaks; native delta logs that stamped literal zero
start hearing truth. JS twins accept the parameter with parity notes.
Pins: provider doubles capture and assert nonzero monotonic generations
across add/update/remove on both surfaces.
The two-implementation contract surface as one pure module (no I/O):
segment header v2 (formatVersion 2 + sealSize in the reserved bytes),
per-record [type u8, version u8] envelope killing the unknown-kind
misclassification trap, the 12-type registry (after-images with minted
ints, tombstones, batch.meta, embed.pending/landed, blob.manifest,
projection.note, bootstrap.baseline, log.genesis with id-space width and
TYPED width-mismatch refusal), vectorLeg inline|{sameAsGeneration} with
writer-enforced single-hop, sector-sealed groups with pad frames, torn-tail
discipline, and GOLDEN BYTE VECTORS pinned so a second (native) reader
implementation can conform byte-for-byte. 50 format pins + a
fault-injecting storage wrapper (tear/drop-sync/fail-append) with 13
self-tests. v1 segments remain readable; nothing writes v2 yet — the
live-format cutover is its own commit.
The storage-authority adoption path, guarded shape: the canonical tree
stays authoritative by default ('tree'); a brain flips to 'log' only
through the verification oracle, and the flip is stored, per-brain,
checked at open only.
- FactLog.ensureSynced(): classic group commit — concurrent writers
append, then join ONE covering fsync (running + queued slots give the
covering guarantee: the sync a caller awaits always starts after its
append landed). Solo writer = immediate sync.
- GenerationStore.logDurability 'deferred' (default, byte-identical to
today: fact durability rides the group-commit flush, ack latency
unchanged) | 'at-ack' (log-authority mode: every single-op ack awaits a
covering log fsync — an acked write's fact survives power loss, by
contract). transact() was already durable-at-return in both modes.
- src/db/logAuthority.ts: the stored switch artifact
(_system/log-authority.json, absent = tree), readLogAuthority, and the
VERIFICATION ORACLE — replay the fact log, fold latest state per id
(digests, never bodies — memory-bounded), diff against the canonical
tree paged; verdict green iff every canonical row is exactly reproduced
AND the log claims nothing canonical denies. Divergences are NAMED by
class (pre-log-record → needs baseline backfill; state-differs;
log-live-canonical-absent; log-tombstone-canonical-present). The flip
REFUSES on red with the first divergence and the cure in the message.
- Brainy: authority read at open (log → durable-at-ack enabled);
logAuthority() / verifyLogAuthority() / adoptLogAuthority() public API.
Nothing flips by itself; nothing changes for existing brains.
A3 of the service-class pair (BRAINY-PROD-LATENCY-TRIAD): a VFS file write
ran the embedder synchronously while the caller waited — 5.6s p50 / 21.4s
p95 per small file on a production deployment, the dominant stage of every
capture write.
- add()/update() gain deferEmbedding: the write acks at durability (data +
metadata persisted, a DURABLE pending marker under
_system/pending_embeds/<id> written BEFORE the commit — orphan-safe
direction); the single-flight background worker embeds the CURRENT data
and swaps the vector in ATOMICALLY (ReplaceInVectorIndex — the row is
never absent from search; a deferred UPDATE keeps serving the OLD vector,
stale-beats-absent per the flicker law). Typed refusals: defer+vector,
defer-without-data.
- CRASH-SAFE: markers are recovered at open by a BOUNDED prefix listing
(never a store walk) and the worker resumes in the background — a crash
can delay a vector, never lose one. A wedged embedder trips a LOUD 60s
hang guard and the worker moves on (marker retained for retry).
- The honest gauges: getIndexStatus().pendingEmbeds + pendingEmbedCount();
awaitPendingEmbeds() is the eventual-vector-index BARRIER for callers
and tests that need searchability before proceeding.
- VFS adopts it everywhere a write path could wait on the embedder:
writeFile (both branches) and directory creation. Pinned in the
strongest form: writeFile resolves while the embedder HANGS FOREVER.
Pins: deferred-embedding 5/5 (ack law · stale-beats-absent · crash
recovery across sessions · VFS hung-embedder ack · typed refusals).
Gates: unit 1928/1928 · integration 765 · conformance 27/27.
DP6/DP8 of the Path Registry (BRAINY-PROD-LATENCY-TRIAD, the proven flicker
mechanism): update paths staged RemoveFromVectorIndex then AddToVectorIndex
as two separately-awaited transaction ops — between them a live row was in
NEITHER index (dark to semantic recall, fine in metadata list). The native
pair widened that window to seconds in production before their side's
visibility-commit fix; the structural cure lands here:
- hnswIndex.updateItem: absent → add; SAME vector → pure no-op (the
production shape — a type-only update re-indexed an unchanged vector,
remove+add did pure damage); changed vector → the node NEVER leaves the
index: synchronous vector swap first (every query from that instant sees
correct distances), then unlink/relink at the node's existing level via
shared internals (linkNode/unlinkNodeEdges refactored out of add/remove;
entry point and maxLevel provably unchanged).
- ReplaceInVectorIndexOperation: ONE transaction leg; feature-detects
provider updateItem (native seam flagged — their side ships updateItem,
then the adjacent remove+add fallback is dead code). Both update staging
sites swapped; delete sites untouched.
- LAZY-OPEN GATE (fleet adoption find, SELF-ENGINE-PAIR-STANDARD): under
disableAutoRebuild, ensureIndexesLoaded assessed ONLY the vector index —
a not-ready native METADATA provider never blocked the completion latch
and every find() silently returned [] on a populated store. All three
providers now vote; any not-ready report falls through to the rebuild.
- docs/path-registry.md: brainy's twin table for the 32 shared path IDs —
service class, budgets, lifecycle, narration, and the cited pin per row;
owed rows named (LC4 doors-open migration, MT4 yielding heals, LC7
downgrade contract) per the lifecycle-sprint choreography.
Pins: update-item-atomic 9/9 (visibility-atomic swap, reverse-index parity
vs fresh rebuild, entry-point invariants) · lazy-notready-honor 2/2.
Gates: unit 1928/1928 (148 files) · integration 760 · conformance 27/27.
A4 of the service-class pair (SELF-ENGINE-LIFECYCLE-SPRINT, David-directed:
'why do we need manual flushes at all?'). The production disease: 829
caller-scheduled per-write flushes convoying into 45-66s write walls —
cadence hand-rolled a layer above the only layer that can see dirty-node
counts and IO pressure.
- BrainyConfig.persistence: policy 'auto' (DEFAULT) | 'manual', with
flushEveryWrites (512) / flushIntervalMs (30s) / flushOnIdleMs (2s)
triggers. Auto = the engine kicks ONE single-flight BACKGROUND flush at
a threshold or when the store goes quiet; write acks NEVER await it (a
hung flush cannot block a write — pinned); a failed background flush is
LOUD and re-arms the trigger. 'manual' restores caller-owned cadence.
- Triggers wired at both write chokepoints (single-op post-commit +
transact post-commit); idle timer unref'd; close() tears the timer down
and drains the flight before its own final flush.
- RECOVERY SEMANTICS documented on the config: canonical records are
durable per-write regardless of policy — a crash between background
flushes loses derived state only, which converges at next open (epoch
machinery + the new incremental aggregation catch-up), bounded by the
un-flushed window. Never data loss.
Pins: write-count trigger fires one background flush with zero caller
calls · idle trigger · manual never self-flushes · THE ACK LAW (writes
acknowledge under a never-resolving flush). Gates: unit 1917/1917 ·
integration 760 · conformance 27/27 — green WITH auto as the default.
SELF-ENGINE-LIFECYCLE-SPRINT + BRAINY-PROD-LATENCY-TRIAD, the four asks:
(a) brain.flush() persists aggregation state stamped at the committed
generation. The stamp used to advance only at close(), so a long-lived
writer that flushes but never closes — the primary production shape —
left every write window behind the stamp, and ANY unclean exit forced a
whole-store backfill walk (per-entity work, measured >60s and
door-starving on a 9k-row production brain) on the first stats call.
(b) BEHIND-stamp adoption becomes adopt + INCREMENTAL CATCH-UP: the exact
missing window (stamp, committed] resolves its affected-id set from the
fact log and reconciles each entity with time-travel before/after reads
(asOf at both window bounds) through the same delta algebra the live
hooks use — cost bounded by writes since the last flush, never store
size, and exact under interleaving because reconciliation targets the
FIXED window end while later writes chain through hooks. Oversized
windows (>5000 affected) and unreadable windows demote to the announced
rescan — never a silent partial serve.
(c) The native provider's parallel rebuildAggregate — on the contract since
8.x but never invoked anywhere — is now the backfill walk's preferred
door: one call per aggregate with source-matched entities, replacing
the per-entity FFI stream.
(d) A delete whose before-image is unavailable can no longer SKIP the
aggregation hook silently (counts drifted upward forever): both delete
paths (remove() and transact) flag an exact rescan, loudly.
Pins: integration (flush stamp; unclean-exit reopen → exact counts through
an add + group-move + delete window with the walk spy proving ZERO
whole-store walks) + unit (provider rebuild invoked once with filtered
entities; flagAllForRescan; reconcile delta algebra). Gates: unit 1913/1913
· integration 760 · conformance 27/27.
BRAINY-PROD-LATENCY-TRIAD Track A1 (David-approved plan): the sort path's
value resolution goes BATCHED — one chunked metadata-record batch pass
serves any N, replacing the serial per-row getNoun loop (62-98ms x 3,224
rows = the measured 199-317 second silent scan on self prod). The
metadata record carries every sortable value: system scalars EXACT
(bucketed-index precision loss can never force a per-row disk read
again) and the user bag via the shape-aware split, both record eras.
- resolveOrderValuesBatch: the one sanctioned value source for ordered
reads (batch door: getNounMetadataBatch -> getMetadataBatch -> chunked
parallel; never serial).
- Column top-K page re-sort and the no-column fallback both rewired.
- B2 down-payment: the no-column fallback ANNOUNCES itself once per
field past 500 rows - silent degradation is illegal.
- THE CALL-SHAPE PIN (tests/unit/utils/metadataIndex-sort-callshape):
zero vector-record reads, batch calls only, latency-blind so it holds
on any machine - the serial loop cannot quietly return. Ordering
contract re-pinned through the batch path (nulls last both directions,
ties by id, never drop).
(! = perf contract change only; no API change. Gates: unit 1904/1904,
integration 758, conformance 27/27.)
The write side of the law, ruled 2026-08-03: data is either in main space
where developers can use anything, or it is in system.*.
- The reserved-name write door DIES: add/update/relate/updateRelation
metadata bags accept EVERY name (confidence, type, id, data, level,
content, ...) as ordinary user fields — indexed, filterable, sortable,
aggregatable, identical to any other field. The remap/enforce/warn
machinery, the reservedFieldPolicy config (now a typed init refusal),
and the compile-time metadata key bans are all removed. The one write
refusal left: keys spelled 'system.*' (namespace forgery), now enforced
on all four write doors.
- STORED RECORDS GO NESTED (v2): engine fields top-level, the user bag
nested verbatim under 'metadata', sealed by a format stamp — by-name
storage discrimination is unsound once colliders are admitted. Legacy
flat records stay readable forever through the shape-aware splitters
(sound for them: the old door refused colliders). Time travel rides the
same split (generation store snapshots whole records).
- Name-based index exclusions DIE: user frame indexes every name; the
excludeFields/indexedFields knobs and their silent-[] holes are gone;
bulk-payload protection is value-shape only, uniform across names.
- Consumer-sweep findings fixed in the same wave: per-type counts read
the frozen 'system.type' column (addToIndex sort, affinity tracking,
cold-count rehydration, VFS type bitmaps — legacy 'noun' fallback for
pre-rebuild reads); resolveHiddenIds addresses 'system.visibility'
(bare 'visibility' was a silent no-op under the law — VFS/system
entities leaked into default reads).
- Fidelity fallout fixed in the owning layers: readEntityFieldAddress
reads the bag first (colliders were absent-shadowed by its own guard)
and never serves system addresses from the bag; blob history refs read
the bag shape-aware; migration transforms now receive ONE normalized
view (engine fields + nested bag) regardless of stored era, and stray
flat-habit keys refuse with the fix in the message.
- THE REOPEN-COLLIDER CONFORMANCE CASE (required before any RC counts as
gates-green): all ten collider names + plumbing names written as user
fields, verified verbatim + queryable across live reads, flush+reopen,
a forced epoch rebuild, and asOf time travel; relation mirror; forgery
refusals; legacy flat-record compat. 8/8 green.
Gates: unit 1901/1901 (exit 0) · integration 758 (exit 0) · conformance
27/27 (exit 0) · consumer test sweep migrated (10 files).
Also completes the v8.10.2 write-granularity law for the transact() plan
path: a metadata-only batch update never rewrites the vector-bearing noun
record (planUpdate staged the unconditional save the update() fix removed).
Seven pins in tests/integration/level-field-shadow.test.ts including the
reporting consumer's exact repro rows; orderBy JSDoc documents the ordering
contract and the announced field-addressing law.
Also: idle PathResolver stats tick no longer logs NaN% every minute (logs
only on new traffic, via prodLog); graph-lsm-* key family recognized as
system resources (kills the per-boot unknown-key warning on provider-backed
brains). Four regression pins in tests/integration/update-write-granularity.
A production deployment's warm report showed metadata: 'unavailable' under a
native metadata provider. brain.warm()'s metadata leg only duck-typed the
built-in JS manager's hydrateAll() method, which a native provider has no
reason to implement.
- MetadataIndexProvider (src/plugin.ts) gains an optional warm?(): Promise<void>
hook, mirroring the existing vector and graph provider hooks. brain.warm()
now checks the active provider's own warm() FIRST, falls back to the JS
manager's hydrateAll() when absent, and reports 'unavailable' only when
neither exists -- never init() as a stand-in, since a native provider's
init() may be a cheap verify rather than a real warm.
- Tests (tests/unit/brainy/warm.test.ts): a live provider instance shaped to
have warm() reports 'warmed' and the hook called with no hydrateAll
fallback; shaped to have neither hook reports 'unavailable' (pins the
honest branch); the unmodified built-in JS manager still reports 'warmed'
via hydrateAll(), unchanged.
Additive scope agreed mid-flight with the native-provider team: a
maintenance-debt observability seam so an operator sees a grind coming
instead of discovering it as a CPU storm.
- New optional maintenanceDebt?(): Promise<ProviderMaintenanceDebt> hook on
all three provider contracts (vector, metadata, graph -- the same three
warm?() lives on). ProviderMaintenanceDebt is fields-all-optional: a
provider reports only what it truly measures (pendingBytes, pendingItems,
lastPassCompletedAt, lastPassOutcome, converging), never an estimate
dressed as fact.
- New public brain.maintenanceDebt(): a pure passthrough -- for each surface
it calls only the active provider's own hook and reports the payload
verbatim, or 'unavailable' when absent. No thresholds, no polling, no
JS-side estimation; the provider owns the numbers, the operator owns the
policy.
- ProviderMaintenanceDebt, MaintenanceDebtReport, and MaintenanceDebtOutcome
are exported from the package root.
- Tests (tests/unit/brainy/maintenance-debt.test.ts): hook present reports
'reported' with the exact payload passed through; hook absent reports
'unavailable' on every surface; mixed surfaces resolve independently of
each other.
RELEASES.md gains the 8.10.1 entry covering both fixes above and this
feature, including the no-hot-retry contract from the prior commit.
A production incident: a native-provider op ground 38-40s inside a transaction,
blew the apply-phase budget, rolled back, and a downstream pipeline hot-retried
the identical operation into a 6-minute CPU storm. Brainy itself never
auto-retried the timeout; the gap was that TransactionTimeoutError only said
"retryable" in prose, with nothing machine-readable for a caller to branch on.
- TransactionTimeoutError gains two typed, always-true fields: retryable
(a later attempt may succeed once the slowness resolves or the budget is
raised) and hotRetryUnsafe (an immediate identical retry re-pays the full
cost that just timed out and can cascade into a CPU storm -- callers must
latch and back off, never loop). context's existing telemetry fields
(timeoutMs, operationIndex, elapsedMs, totalOperations, operationName) are
now documented as the caller's backoff inputs.
- Updated the "retryable" doc-prose sites (transact()'s timeoutMs option,
transactionBudgetFloorMs, Transaction.execute()'s contract) to point at
the new fields instead of bare prose.
- Regression pin (tests/unit/transaction/timeout-never-internally-retried.test.ts):
an execution counter proves the engine never re-drives a timed-out
operation, through both the single-op engine TransactionManager/Transaction
drives for every single-record write, and add()'s upsert-race retry loop
(which must exit on the first TransactionTimeoutError, never treat it like
the lost-insert-race signal it retries on).
- Removed TransactionManager.executeTransactionWithResult -- zero callers
anywhere in the codebase.
Ruled today: the project's one public home is source.soulcraft.com. The
old public repo is archived history and no longer part of any release.
- package.json repository/homepage/bugs now point at the forge (this is
what the npm page links as Repository/Homepage/Issues)
- README CI badge reads the forge pipeline; CONTRIBUTING drops the
mirror paragraph (forge account or email patch were already the ruled
contribution paths)
- release.sh: mirror push + external release step removed; publishes go
forge-first (box-held write token, temp userconfig so the token never
hits argv; a forge-publish failure aborts before the storefront so the
pair can never diverge), then npmjs with the scope-override pin (the
fleet npmrc maps @soulcraft to the forge and scope mappings beat
--registry); release page created via forge API when a token is
present, loud skip otherwise; changelog compare links point home
- dead external CI workflow removed (.forgejo/workflows/ci.yml is the
live pipeline)
Historical CHANGELOG links to the archive stay as written - history is
history and the archive serves them read-only.
Ruled today: the project's one public home is source.soulcraft.com. The
old public repo is archived history and no longer part of any release.
- package.json repository/homepage/bugs now point at the forge (this is
what the npm page links as Repository/Homepage/Issues)
- README CI badge reads the forge pipeline; CONTRIBUTING drops the
mirror paragraph (forge account or email patch were already the ruled
contribution paths)
- release.sh: mirror push + external release step removed; publishes go
forge-first (box-held write token, temp userconfig so the token never
hits argv; a forge-publish failure aborts before the storefront so the
pair can never diverge), then npmjs with the scope-override pin (the
fleet npmrc maps @soulcraft to the forge and scope mappings beat
--registry); release page created via forge API when a token is
present, loud skip otherwise; changelog compare links point home
- dead external CI workflow removed (.forgejo/workflows/ci.yml is the
live pipeline)
Historical CHANGELOG links to the archive stay as written - history is
history and the archive serves them read-only.
Origin moved to the private forge; the public repo is now a mirror with an
unknown sync cadence. The release flow creates the GitHub release directly,
and a missing tag there would be silently created at the default-branch
head - the wrong commit. The script now pushes branch+tag to the mirror
itself and hard-stops if the tag's commit on the mirror differs from local,
before anything irreversible (npm publish) happens.
Reconciles the vector-index rename to the ruled three-layer naming: the
provider contract's identity field is now a REQUIRED readonly name (was
optional providerId), self-reported and truthful, rendered as
[vector-index:<name>] where the index identifies itself and stamped into
the op-name strings journals already parse (AddToVectorIndex(<name>)).
The built-in JS engine names itself hnsw-js. A runtime provider instance
compiled against the previous optional contract is tolerated - never
crashed on, never silently mislabeled: it stamps unknown-provider and
emits one loud warning naming the missing field. Graph index operations
keep their static names (they never interpolate provider identity), and
no public API exports a provider-routed hnsw-carrying name, so no
deprecation shim is required.
Three pieces addressing the cold-restart-write incident where a production
deployment's first writes after every restart (33-35s each on a cold page
cache) blew the op-count-scaled transact budget mid-batch: every write is
itself a multi-op transaction, so one cold operation consumed the whole
budget, the gate before the next operation tripped, and the write rolled
back atomically - refused, retried, and refused again until the page cache
warmed passively.
- The budget's start-gating contract is now explicit and pinned: it gates
STARTING the next operation, never rolling back completed work for
elapsed time (the shipped schedule since 8.7.0, now stated in contract
JSDoc, guarded by code for operation 0, and enforced by regression
tests). The 30s floor is configurable via transactionBudgetFloorMs for
stores whose cold operations legitimately run long.
- New brain.warm() eagerly loads the vector index, metadata index, and
graph adjacency so first operations after a cold restart run at
steady-state cost. Returns a WarmReport with an honest per-surface
outcome (warmed / probed / unavailable) - never reports a probe as a
warm. warmOnOpen: true runs it during init(). New optional provider hook
warm() on the vector and graph plugin contracts.
- Vector-index transaction op classes renamed from the backend-specific
AddToHNSWOperation / RemoveFromHNSWOperation to backend-neutral
AddToVectorIndexOperation / RemoveFromVectorIndexOperation, stamping the
active backend into the emitted op-name string (AddToVectorIndex(js-hnsw)
vs a native provider's own identity) so journals never misdirect an
operator toward an index that isn't running.
The packed tier goes live inside GenerationStore:
- Two-tier reads: getDelta / readBeforeImage / readGenerationRecords
fall through live-tier → sealed segments (live-tier-wins: a crash
mid-fold leaves a duplicate representation, never a gap). Cold-open
seeds committedRanges from the segment manifest via interval merge —
packed generations resolve without their directories existing.
- repackHistory({timeBudgetMs, batchGenerations}): folds cold
generations (older than the newest 1024) oldest-first into sealed
segments, deleting per-generation directories only after segment +
manifest are durable. Public API + automatic time-bounded pass at
close() (before compaction, so reclaim can drop whole segments);
re-representation only — the sole history transform under the
archival profile. Early stop = consistent prefix, next pass resumes.
- compact(): packed generations reclaim logically in the loop and
physically at whole-segment boundaries via dropSegmentsBelow (the
frozen partial-segments-wait rule).
- generationDigest(g) (D8): deterministic content digest through g —
sealed-segment checksum chain + live-tier delta hashes; O(segments +
live window); RangeError out of range, GenerationCompactedError
below the horizon (a gate can never silently pin reclaimed history).
Four end-to-end pins: asOf answers byte-identical across fold + cold
reopen with folded dirs physically gone; repack+reclaim composition;
digest reopen-stability/divergence/loud-horizon; budget no-op+resume.
First stage of the co-frozen D1+D3+repacking unit: the format core,
self-contained under _generations/segments/.
- seg-<firstGen,20pad>.bgs: append-once packs of consecutive
generations (magic BGS1; frame = u32 len + u32 crc32c + msgpack
[generation, timestamp, delta, records, flags]; flags reserves
compressed-payload evolution without a format break). Sealed
segments are immutable — fold refuses overlap with sealed ranges.
- seg-<firstGen>.idx: DERIVED sidecar (per-generation frame offsets +
per-id generation postings + checksums); lost/corrupt sidecars
rebuild from their segment loudly; a damaged segment (frame CRC
mismatch) fails loudly, never serves wrong bytes.
- manifest.json: the one discovery path — open() reads it and never
lists the packed backlog (the scan-wedge class's cure); refuses a
newer manifest version rather than serving partial history.
- D3 semantics: dropSegmentsBelow reclaims WHOLE segments at
boundaries only and bumps compactedBelow durably; archival-profile
enforcement stays with the caller per the co-freeze.
- D8 rider: digestThroughPacked(g) — deterministic crc32c chain over
sealed-segment checksums (+ frame-level prefix mid-segment),
O(segments), reopen-stable.
Also fixes a cross-adapter contract bug the suite caught: memory
storage's deleteObjectFromPath ignored the raw-bytes store, so
deleteRawObject on a raw-bytes path (fact-log or segment files)
silently no-op'd — deletes now match filesystem unlink semantics.
Six pins. Wiring into GenerationStore (two-tier reads, the repacker,
cold-open manifest path) lands with the rest of the unit before its
release; cortex's fact-record/stamp shapes reconcile the sidecar
keying when they post.
Stage-2 D1 contract item (co-frozen): a fact scan may be slow, never
silent. batches() now races its FIRST pull against
SCANFACTS_FIRST_BATCH_MS (10s, exported; test-overridable) — a wedged
or unreadably slow store produces a loud abort naming the contract
instead of a consumer hanging indistinguishably from progress (the
production shape: a heal against a generations-backlogged brain
wedged silently on the first segment read).
Only the first pull is raced: the bound is time-to-first-batch (proof
the producer is alive), not per-batch pacing, and it runs only while
a pull is pending — consumer think-time between pulls never counts
against the producer (pinned).
Three pins: wedged-store loud failure within the bound, healthy scan
untouched end-to-end, slow-consumer immunity.
First edition of the per-release performance-envelope contract: every
number measured against the built dist on stated hardware, never
projected. Sub-0.1ms get/related (adjacency O(degree), scale-flat),
1-9ms indexed metadata finds, ~178ms semantic (query embedding
dominates), ~167ms durability-priced single-op writes flat across
scale, 8-45ms steady-state flush independent of history backlog
(the 8.9.0 change). Two weak spots stated honestly: addMany commits
per-item today (batched chunk commits belong to the unified-commit
roadmap), and pure-JS warm open grows with corpus (4.9s at 10k) —
the native accelerator's reason to exist. Refresh rule: any release
touching a measured path re-measures in the same release.
clearInterval() stops future heartbeat ticks but not one already in
flight: a straggler tick past its ownership guards could land its
atomic lock rewrite AFTER releaseWriterLock()'s unlink, re-creating
the lock file as a phantom that blocks the next writer until the
stale TTL expires (~60s) — the pool-eviction reopen case. Found as an
ENOENT heartbeat warning during benchmark teardown; the quiet variant
is the harmful one.
releaseWriterLock() now awaits the in-flight tick (tracked per tick,
self-clearing) before reading/unlinking, so a straggler's write always
lands BEFORE the unlink and gets removed with everything else. The
heartbeat's ENOENT is also now benign-by-contract (lock or directory
removed under us — the next acquire recreates it); other errors stay
loud.
Pin: straggler-past-guards simulation — lock file absent after close,
directory immediately claimable (fails on the undrained code).
flush() is durability work: it must cost what the current window's
deltas cost, never what the history backlog costs. Under adaptive
retention the byte budget derives from free memory, so bulk-load
pressure shrank the budget exactly at peak write volume and flush paid
actual reclaim inline — a production deployment measured single writes
blocked 25-191s behind reclaim-on-flush.
- flush() no longer calls autoCompactHistory(); close() is THE
auto-compaction site (already ran there; now alone).
- Every auto pass is time-bounded (CLOSE_COMPACTION_BUDGET_MS = 5s):
reclamation is oldest-first, so an early stop is a consistent prefix
and the next pass resumes. Explicit compactHistory() gains an
optional timeBudgetMs for caller-chosen maintenance windows.
- Documented trade stated where operators read: a long-lived writer
that never closes accumulates history until its next explicit
compactHistory() — predictable writes, explicit maintenance.
Pins: flush-never-reclaims + close-reclaims-durably (db-mvcc), bounded
pass stops-then-resumes as a consistent prefix (generationStore unit).
Four fixes from a consumer conformance report, one root disease — two
field-resolution regimes where there must be one:
- The delete/update aggregation hooks fed the engine a partial entity
view (type/service/data/metadata only), so a reserved-field groupBy
(subtype, visibility, ...) resolved to a nonexistent group on the way
down: counts drifted upward forever after deletes, and updates moving
an entity between reserved-field groups double-counted. The hooks now
pass the full-fidelity view via entityForAggFromRawRecord (every
reserved field top-level, mirroring the add path); the update sites
pass the full get() view instead of a hand-rolled subset.
- Aggregation source.where resolved fields only against the custom
metadata bag, so where on a reserved field silently matched nothing.
The matcher now resolves each filtered field through
resolveEntityField — the same single source of truth groupBy uses.
- removeMany() with no usable selector (bare array passed positionally,
empty params, ids: []) resolved successfully having deleted nothing.
All three now throw; the two legacy tests that pinned the silent
no-op as 'graceful' now pin the refusal.
- find() where keys accept both spellings: a metadata.-prefixed key
falls back to its flattened spelling when the prefixed one is not
indexed (metadata is flattened at index time). A literal nested custom
key named metadata still wins when indexed as spelled.
Five regression pins in aggregate-reserved-fields.test.ts (4 of 5 vary
red on the unfixed code).
scripts/push-docs.js collects docs/**/*.md with public:true frontmatter
and POSTs them (batches of 10, idempotent per slug) to the docs ingest
door after npm publish — release.sh step 12. Absent secret = loud skip
(publish already happened; the serving side can interim-sync); a failed
push exits non-zero so the docs site never silently trails npm. The
combined /docs landing index is deliberately NOT pushed per-repo — it
spans both engine corpora and is authored on the serving side.
Under default adaptive retention, every flush() recomputed total history
bytes by walking EVERY committed generation's delta — O(all generations)
with disk re-reads past the 4096-entry delta-cache bound. On a production
brain with 70,000+ accumulated generations this turned every write into a
full-tail scan (60-100s writes, escalating with history growth), even
though the free-RAM budget never tripped and nothing was ever reclaimed
(SELF-GENERATIONS-GROWTH).
historyBytes() now maintains a running total: seeded by one walk on first
use, then updated incrementally at both commit paths (+bytes) and the
compaction reclaim loop (−bytes), dropped on reopenAfterRestore. The
adaptive retention check on every flush is O(1). Invariant regression-
pinned: running total ≡ fresh walk through transact commits, single-op
group commits, and compaction.
New brain.historyStats() (exported HistoryStats): read-only generation
count / bytes / generation+timestamp range / horizon / retention mode /
effective budget — the one-call per-brain fleet audit for retention
exposure.
The post-import background deduplication pass (a merge-DELETE writer,
debounced ~5 minutes after import) had four lifecycle defects:
- enableDeduplication: false did not gate the background schedule — an
import that explicitly opted out could still have entities auto-removed
minutes later. The flag now gates both the inline and background passes.
- Each import() constructed its own coordinator-owned deduplicator, so the
debounce never spanned imports (N imports = N delete timers). The brain
now owns a single lazy instance (getBackgroundDeduplicator).
- close() never cancelled pending dedup; a delete pass could fire against
a closed brain. close() now cancels it first.
- The 5-minute timer held the process open (exit-hang class); now unref'd.
Four regression tests pin the contract (background-dedup-lifecycle);
import guides document that false disables both passes.
- New src/utils/osLimits.ts: reads RLIMIT_NOFILE (soft/hard, from
/proc/self/limits) and vm.max_map_count at open — once per process,
Linux-only, measurement-only — and warns loudly when either sits below the
pool-scale floors (soft NOFILE < 65536, max_map_count < 262144), with the
exact raise commands. On stock defaults the failure otherwise arrives as
EMFILE or a failed mmap deep inside an index open, long after the cause
stopped being visible. An unreadable limit produces NO warning — no
measurement, no claim — so non-Linux platforms stay silent.
- Exported for ops doors: checkOsLimits() returns the full OsLimitsReport;
floors exported as constants. Wired fire-and-forget in performInit after
storage init; the check can never affect open.
- Unit tests pin the parser (incl. 'unlimited'), the floor thresholds, the
null-never-warns rule, and the off-Linux silent path.
- acquireWriterLock now CLAIMS with an atomic create-exclusive write (O_EXCL)
inside a bounded retry loop: two processes racing an absent lock can never
both succeed (the old read-then-tmp-rename flow let the loser keep running
unlocked, silently). An EEXIST loser re-evaluates and either throws loudly
with the winner's details or performs a verified stale-takeover (re-read
before unlink so a lock that changed hands mid-deliberation is never
clobbered). Exhausted contention fails loudly instead of degrading into a
lockless open.
- BRAINY_WRITER_LOCKED passes through init() unwrapped: the error documents a
machine-readable contract (err.code + err.lockInfo with the holder's
pid/host/heartbeat), but init's blanket error wrapping stripped both,
leaving consumers a message to regex against.
- Two contract tests added: stale-foreign takeover installs OUR lock via the
atomic claim; the conflict error carries code + lockInfo at the public
init() surface.
- The transact apply budget scales with the batch — max(30s, opCount x 2s) —
or is exactly the caller's new TransactOptions.timeoutMs. A flat 30s cap
silently limited honest bulk work to ~15 operations on network-attached
storage (~2s/op measured in a production import) while looking generous for
small batches. Internal batch paths (removeMany chunks) get the same
scaling via the shared transactTimeoutBudget helper.
- TransactionTimeoutError now reports the operation it stopped at as i/N with
the operation's name, elapsed vs budgeted time, and states the batch rolled
back atomically and is retryable; context carries the fields
programmatically.
- The transact operational envelope is documented (optimistic-concurrency
guide): budget math, chunking with ifAbsent idempotency, when to reach for
addMany vs transact, and the precompute pattern (embedBatch + per-op
vector) that keeps inference out of the commit path.
- Embedding claims made honest per an end-to-end probe of the built dist:
batch and single embedding paths produce bit-identical vectors and a
vector-supplied add is fully searchable, but batch throughput on the
default WASM engine measures comparable to sequential (~160ms/text) — the
5-10x speedup claim in the addMany JSDoc is replaced with the measured
reality and the actual win (inference outside the budgeted write path).
- New public API: walks every canonical relationship record, queries the same
read path applications use (related() with all visibility tiers included),
and classifies every discrepancy into its failure family: missingFromReads
(records the read path omits — stale adjacency), danglingEndpoints
(endpoint entity gone — the partial-delete scar class), readOnlyVerbIds
(read-path edges with no canonical record — ghosts). Design-hidden
internal/system edges are counted separately so intentional hiding is never
misclassified as loss. Counts exact; example lists capped with an explicit
truncatedExamples flag; loud narration on incoherence; mutates nothing.
- Classification core lives in src/graph/graphAudit.ts behind injected seams
(canonical walks + the end-to-end read) so the discrepancy logic is testable
in isolation; brainy wires the seams to storage walks and related().
- getNounIdsWithPagination now refuses an undecodable resume cursor loudly —
the third and final pagination walk brought under the 8.5.2 cursor contract.
- Types exported: GraphAuditReport, GraphAuditDiscrepancy. Docs: the
inspection guide gains the audit -> repair -> audit verification ritual.
- Aggregation backfill walks build into a STAGING map and swap in atomically
on completion. A mid-walk failure drops the staging map — the previous live
state keeps serving, the aggregate stays flagged pending, and the storage
error surfaces to the failing query. Previously the walk wiped live state
before a scan that could throw, never cleared the pending flag on failure,
and re-ran a full walk on every subsequent query: a silent wipe/walk/throw
loop at the caller's retry rate.
- Failed walks are latched: retries within a 30s cooldown rethrow the recorded
error instantly instead of re-walking, so a tight caller-side retry loop
costs one loud error per query, never a full store walk per query.
- Persisted aggregation state is stamped with the store's committed generation
at flush; reopen adoption requires stamp equality. Stale state (unclean
shutdown) or over-counting state (a log truncation on a copied store pulled
the watermark back) triggers exactly one loud rescan, never a silent adopt.
- The backfill/adoption path narrates: adoption decisions, walk start/finish
with counts and duration, and failures all log by default.
- getNouns/getVerbs refuse a supplied-but-undecodable pagination cursor with a
loud error instead of silently restarting the walk at offset 0 (which
re-served page 1 forever to any while(hasMore) caller).
- The graph cold-load verb walk aborts loudly on a missing or non-advancing
cursor with hasMore=true.
- A versioned index provider whose generation is AHEAD of the committed
watermark (torn copy / crash-recovery truncation) is now named loudly at
open, alongside the existing behind-direction message.
- AggregationIndex: defineAggregate before init no longer forces a backfill.
init reconciles instead of clobbering: the app definition wins, persisted
state is adopted on hash match, a write landing pre-adoption forces an exact
rescan, and a successful state load clears the backfill flag. New ready()
settles every adoption decision before query paths consult backfill state.
- brainy: backfills are single-flight and batched. Concurrent queries share one
store walk and every pending aggregate fills from that same walk; the old
behavior let each concurrent query wipe the others' partial state and start
its own full walk, so a store under steady aggregate traffic never converged.
queryAggregate also waits for persisted definitions before deciding an
aggregate does not exist.
- paramValidation: recordQuery is telemetry-only. The duration-based cap
ratchet (x0.8 per recorded query while lifetime-average exceeded 1s, floored
at 1000 - below the documented 10000 auto floor, reported under a stale
basis label) is removed; the cap comes from its construction-time basis or
explicit overrides alone.
- storage: __aggregation_* and singleton system keys (brainy:entityIdMapper)
are recognized before the unknown-key warning fires; routing is unchanged.
- docs: find-limits cap-immutability note, aggregation reopen/adoption
semantics, RELEASES.md 8.5.1 entry.
- New public guide (guides/external-backups): the sparse-mmap reality
for external backup tooling — why a brain directory can show 100+ GB
apparent size on a small disk, which files are sparse, tar czSf /
rsync --sparse / cp --sparse=always, live-store caveats, and what
persist()/restore() already handle natively.
- New public concept (concepts/generation-fact-log): what facts are
(after-image commit records, body-less tombstones), the crash-safety
model (checksummed frames, reconcile-to-committed, absent = never
committed), the scanFacts()/factSegmentPaths() surfaces with the
telemetry shape, family stamps + open-time coherence, and the
storage capability seam for plugin authors.
- Cross-link from the snapshots guide; sparse notes added to the
public persist()/restore() JSDoc (the operator-facing sites).
setTimeout can fire a few ms early under load (timer coalescing) — a
10ms sleep measured 9ms and failed the >=10 assertion, tripping a
release gate. Sleep 25ms, assert >=20: the test verifies time is
MEASURED, not the OS timer's precision.
- storage.committedGeneration(): the committed watermark exposed on the
fact-scan capability, so a provider compares its stamp's
sourceGeneration against the store's truth without parsing the
private manifest format (the capability source now carries both the
live fact log and the committed closure).
- Pinned contract tests: fsync-before-ack (transact passes today; the
single-op path is pinned via it.fails as the documented target —
group commit must become LATENCY batching, ack waiting on the shared
fsync, never durability skipping; the pin flips red when that lands
so there is no cliff to discover) and scan-stability-under-rotation
(a scan snapshot yields exactly its facts — no gaps, duplicates, or
bleed-in — while segments rotate beneath it; the reclaim-during-scan
variant lands with fact compaction, which does not exist yet).
- FactLog accepts a rotateBytes option so tests exercise real rotation.
Three additive surfaces for native index providers, which hold only
`storage` and must never construct their own fact-log reader (the log's
open path is writer-side — it reconciles by truncating/rewriting):
- Fact-scan capability on the storage adapter (the getBinaryBlobPath
pattern): the host brain wires a closure over its LIVE fact log at
init; providers call storage.scanFacts()/factLogHeadGeneration()/
factSegmentPaths() and fall back to the enumeration walk on null.
Restore/reopen swaps the underlying log transparently.
- The family-stamp trio (readFamilyStamp/writeFamilyStamp/
verifyFamilyStamp + types) exported from @soulcraft/brainy/internals,
so 'one verifier' is literally one function shared with the native
side, never two synchronized copies.
- Rollup invariants widened to number|string: content fingerprints
(e.g. a per-tree SHA-256) are valid invariant values; strict equality
either way, and a type mismatch reads as incoherence, never a pass.
The canonical entity tree now carries a FAMILY STAMP
(_system/family-stamps/entity-tree.json, JSON — forensics stay
terminal-readable): which committed generation the tree reflects
(sourceGeneration) plus the rollup invariants (entity/relationship
counts) that verify a millions-of-files projection whole where per-file
checks cannot scale. Written at flush/close boundaries; open-time
coherence is a COMPARISON, not a walk:
- coherent / absent (legacy store) → silent
- behind → benign for the tree (it is written BY the commit; only the
stamp is stale after a crash between commit and flush) — refreshes at
the next flush
- incoherent (counts diverged at equal generation, or a stamp AHEAD of
the log head) → loud, names the failing invariant; repairIndex()'s
unconditional recount heals and RE-STAMPS so repair leaves a coherent
stamp behind
- a fault reading the stamp is UNVERIFIABLE — never conflated with
absence
One verifier (verifyFamilyStamp, exported) reads both member modes:
enumerated (exact byte size per member, bounded families) and rollup
(invariants, unbounded families). New exports: readFamilyStamp,
verifyFamilyStamp, ENTITY_TREE_STAMP_PATH, FamilyStamp, StampMembers,
StampVerdict.
Every committed generation now also appends a FACT — an after-image
commit record (what each touched entity/relationship became, or a
body-less tombstone for a removal) — to an append-only, crc32c-framed
segment log under _generations/facts/. The before-image history and the
canonical tree remain authoritative; the fact log gives consumers ONE
sequential, self-verifying stream (index heals, incremental replays)
in place of a per-entity directory walk.
- Wire format: positional msgpack facts [generation, timestamp, ops,
meta, blobHashes]; op = [kind u8, id bin16, record | nil tombstone];
32-byte segment header (magic, formatVersion, firstGeneration,
zeroed+verified reserved); length+crc32c frame per fact; zero-padded
segment names so lexicographic order == generation order; JSON
manifest with an atomic rename flip, manifest-first rotation.
- Commit protocol: facts append+fsync BEFORE the commit point inside
the existing durability window, so a crash can only leave the log
AHEAD of committed truth — open() truncates back (torn tails detected
by CRC). Absent generation = never committed; a scan can never see an
uncommitted fact. transact() facts are durable-on-return; single-op
facts ride the group-commit flush exactly like buffered history. A
fact-append failure fails the write, loudly — a silent gap would be a
lie a later replay discovers.
- New public surface: brain.scanFacts() (sequential batches with heal
telemetry: head/segments/approx up front, per-batch generation range
+ bytes + segment id, loud abort on gaps, summary cross-check) and
brain.factSegmentPaths() (immutable sealed segments for zero-copy
consumers; the mutable tail excluded). Exported types CommitFact,
FactOp, FactScanBatch, FactScanHandle.
- Storage: optional binary raw-byte primitives (appendRawBytes,
readRawBytes, writeRawBytes, rawByteSize) on StorageAdapter —
feature-detected; filesystem + memory adapters implement them; an
adapter without them hosts no fact log. Fact segments are byte-copied
(never hard-linked) into snapshots. The _generations/facts/ namespace
is registered as a protected family (rebuildable: false): no sweeper
or GC may delete under it.
- New crc32c (Castagnoli) utility with RFC known-answer tests.
Ports the fresh-brain probe that closed the type+subtype lens-drop
investigation into the permanent suite: a 7-pair corpus seeded through the
real write API (never restored bytes), every lens checked id-for-id against
an unfiltered canonical scan, warm AND after a cold reopen, plus the
type+subtype update-flip leg. Guards the under-inclusion class the egress
integrity guard structurally cannot catch.
A cross-directory rename added the new parent's Contains edge but skipped
removing the old one (a 'not critical' shortcut), leaving the moved entity
a child of BOTH directories: readdir(oldDir) kept listing it, re-creating
the old path showed the name twice, and tree-walking consumers saw the
file in two places. Reported live by a production deployment (37 stale
containment ghosts; blocks tree-sync integrations).
- rename() now removes the old parent's vfs containment edge(s) by edge
id resolved from the graph's own adjacency (the removal law: removal
never requires reading the removed thing), and a move to the root gets
its containment edge (previously skipped -> orphaned from readdir('/')).
- New vfs.repairContainment(): reconciles every VFS entity's containment
edges against canonical metadata.path — removes stale/duplicate vfs
edges, restores missing ones, never touches user knowledge edges. Wired
into repairIndex() so the one operator ritual heals existing ghosts.
- Regression: tests/integration/vfs-rename-containment.test.ts (7).
Persisted entity totals inflated permanently: a delete's count decrement
was sourced from re-reading the record being removed, so a null read
(replace race, or a ghost left by a pre-8.3.1 partial delete) silently
skipped the decrement while the paired add had counted — and
Math.max(persistedTotal, scanned) meant no disk cleanup could ever
lower the reported total. Reported by a production deployment with a
write→delete→re-create repro minting drift within hours.
- The caller's pre-delete read now rides through the entire delete path
(remove()/removeMany()/plan → DeleteNoun/DeleteVerb operations →
deleteNoun/deleteVerb → deleteNounMetadata/deleteVerbMetadata): a
null internal read falls back to the known prior record instead of
skipping the decrement. StorageAdapter.deleteNoun/deleteVerb gain an
optional priorMetadata parameter (additive).
- rebuildTypeCounts() rebuilt only the type-statistics arrays and
computed the total just to LOG it — the persisted scalar
(counts.json) survived every rebuild. One canonical walk now rebuilds
every counter rollup (scalar totals + per-type maps + type
statistics) and persists them; repairIndex() runs the recount
unconditionally, since counters can be inflated over clean shelves.
Two production-reported spine fixes:
- Canonical delete is a FULL removal: remove()/removeMany() deleted the
metadata leg but never the canonical vectors.json leg or the <id>/
directory, leaving ghost rows that inflated enumerated counts forever,
read as damage scars, and caused duplicate readdir entries for
re-created VFS paths (the stale-unpost path). The delete operation now
routes through storage.deleteNoun — both legs + the entity container —
with a full two-leg before-image rollback; deleteVerb symmetric; the
blind 'no metadata file' catch that masked real faults is gone. The
generation log still holds the delete's before-image, so asOf() history
is unchanged. repairIndex() gains a conservative, loud orphan-prune
sweep (containers with no metadata content leg) + count recompute for
stores damaged by earlier versions.
- Family-scoped migration gate: every read blocked on the whole-brain
migration lock even when the migrating index was irrelevant. The gate
now scopes to the index families a read actually consults — canonical
reads never wait; find() waits only on its query shape's families;
graph traversals wait only on the graph family. Writes keep the
conservative whole-brain wait. A read that needs the migrating family
still blocks (bounded) with the retryable MigrationInProgressError.
Consumer summary of the 8.3.0 minor: 16-way parallel canonical enumeration +
id-only getNounIdsWithPagination (index-heal speedup, standalone); the ADR-004
cross-layer integrity contract (validateInvariants delegation + repairIndex
native rebuild); and the registered-blob family contract (declared index blobs
undeletable). The two contract pieces are inert until a native provider
implements the matching hooks.
The canonical enumeration walk (getNounsWithPagination) hydrated each entity's
vector + metadata ONE-AT-A-TIME inside the shard loop — every enumeration paid
N x per-op-latency serially, and every index heal enumerates canonical, so this
was the dominant multiplier in the measured multi-minute heals (cortex heal-cost
decomposition).
- Hydration is now 16-way bounded-concurrency (matching the wave cor's native
rebuild uses): heal wall-clock becomes ~2xN/16 x per-op instead of N x per-op.
Order, cursor resume (skipped nouns still never read), filters, peek/hasMore
and totalCount are all preserved — pages are byte-identical to the serial walk.
- New getNounIdsWithPagination(): the id-only opt-out for callers that own their
IO schedule (an index heal). Unfiltered = ids straight from the shard paths,
ZERO per-entity reads; filtered hydrates metadata only (16-way). Same
cursor/offset/nextCursor contract, so it is page-compatible with the hydrating
walk.
Regression: pagination-parallel-hydration.test.ts pins paged==big-page identity,
ids==items order, zero-read id-only, and filter parity. 103 storage tests green.
The class-killer for the lost-main.dkann incident (ADR-004 §7). A provider can
declare a derived-index blob FAMILY (a set of members that are load-bearing
together, e.g. vector-base = main.dkann + main.slotmap + main.slotrev). Once
declared:
- deleteBinaryBlob / removeRawPrefix REFUSE to remove a declared member, throwing
the new ProtectedArtifactError — an in-process GC / sweeper is now INCAPABLE of
deleting a load-bearing index file (COLD != DEAD). Intentional retirement is an
explicit unregisterDerivedFamily(name) first.
- The declaration persists to _system/derived-artifacts.json, so protection
survives a reopen; clear() resets it with the rest of the derived footprint.
- checkDerivedFamiliesPresent() names any member missing on open (the catch for an
EXTERNAL deleter that bypasses the in-process refusal) → rebuild from canonical.
- Transients (*.tmp.*, *.rebuild-tmp, *.rotate-tmp) are never protected; a
namespace family protects a growing prefix (seg-*).
New StorageAdapter surface (optional): registerDerivedFamily / unregisterDerivedFamily
/ listDerivedFamilies + DerivedFamilyDeclaration; new exported errors
ProtectedArtifactError / DerivedArtifactMissingError. Enforcement is inert until a
provider declares a family (no regression). Cor declares its 6 families + does the
atomic set-swap in 3.0.15 (M2); brainy builds the contract now. 8 tests.
validateIndexConsistency() was blind to native providers — it only ran the JS
metadata index's own check, so a native provider whose manifest/segments/counts
had diverged still read as "healthy". Per ADR-004 §6 it now feature-detects and
aggregates each provider's optional validateInvariants() (a never-throwing,
<50ms self-report of its own cross-layer invariants), names any failing
invariant with its numbers in the recommendation, and exposes the per-provider
reports. A provider that violates the never-throw contract is surfaced as
unhealthy, never swallowed.
repairIndex() now reconciles NATIVE derived state from canonical too: it consults
each provider's invariants and calls rebuild() on any whose failing invariant
asks for heal:'rebuild' — the native counterpart of detectAndRepairCorruption().
New provider surface: optional validateInvariants(); new exported types
ProviderInvariantReport / InvariantResult / InvariantHeal. Additive, no break.
Cor implements the hook in 3.0.15 (M2); brainy builds against the shape now
(feature-detected, inert until a provider exposes it). 5 tests.
Closes the "dishonest readiness proxy" anti-pattern (Pattern A): size()>0 /
isInitialized were treated as "this index serves queries", but a cold native
index can load its COUNT before its SERVING structure, so a query returned a
silent [] indistinguishable from "no such data". A shared assessIndexReadiness()
now reads only the provider's honest isReady() signal (never size()), applied at
every site:
- Vector: a one-shot verifyVectorLive() guard on the semantic/proximity search
path (a pure semantic find({query}) has no filter, so nothing guarded it). It
prefers isReady(), else a known-vector self-match probe; self-heals via rebuild
or throws the new VectorIndexNotReadyError instead of a silent [].
- Graph: getVerbsBySource/ByTarget skip the fast path when the provider reports
not-ready (falling to the canonical shard scan), plus a one-shot probe that
self-heals a no-isReady provider whose adjacency did not cold-load.
- getIndexStatus(): folds in per-index honest `ready` (making `populated`
honest) + rebuildFailed/rebuildError/degradedIds, so a readiness probe never
200s a brain that is still warming up or degraded.
Unblocked by the native providers now reporting serving-truth (graph via
SSTable-residency readiness, vector via durableBaseLoadFailed). New export:
VectorIndexNotReadyError. getIndexStatus gains additive fields. No breaking API.
13 new tests; existing readiness guards green.
The Pass-1 hardening that makes loadBinaryBlob distinguish genuine absence
(ENOENT -> null) from a real fault (EIO/EACCES/... -> throw) was held out of
8.2.6 because the native accelerator's two column-store read sites still relied
on null-on-error. That accelerator release has now hardened those sites to
handle the throw (a faulted segment read marks the field unavailable and throws
a named error), so the fault-propagating form is restored and ships lockstep.
A present-but-unreadable index blob no longer masquerades as absent (which drove
needless rebuilds / empty reads). Adds the loadBinaryBlob leg to the blob
durability suite (absent -> null; present -> bytes; real fault -> throws).
Consumer-facing summary of the Pass-1 durability + integrity release: durable
blob-write honesty, single-op history durability (PendingFlushDurabilityError),
degraded-index surfacing, full-footprint clear(), count symmetry, and loud
index-maintenance / aggregation failures. loadBinaryBlob fault-propagation is
held for the native lockstep and deliberately omitted here.
The Pass-1 hardening made loadBinaryBlob distinguish genuine absence (ENOENT →
null) from a real fault (EIO/EACCES/… → throw), so a present-but-unreadable blob
stops masquerading as "absent". The native column-store still has two call sites
that rely on the old null-on-error contract; publishing the throw ahead of them
would break a consumer running new brainy + old cortex.
Per the cortex coordination (accept: cor first, then brainy), this temporarily
restores the shipped 8.2.5 swallow-on-fault behavior for loadBinaryBlob ONLY, so
the rest of Pass 1 — saveBinaryBlob write-honesty, clear() native wipe,
pending-flush durability, count symmetry, degraded surfacing, aggregation
loudness — can release now. The method carries an inline restore guide; the
throwing form goes back in and ships lockstep with the native hardening. No other
Pass-1 change depends on this behavior (ColumnStore's own fault test uses a
direct-throwing storage stub), so the split is behaviour-neutral for every
consumer versus 8.2.5.
The debounced materialization caught its failure with an empty `catch (() => {})`,
silently leaving the materialized Measurement entity stale; the aggregation-index
init() state-load failure was swallowed the same way, leaving aggregates reading
empty with no signal. Both are non-fatal (values rebuild via backfill-on-query),
but a silent stale/empty read violates loud-errors-never-quiet-losses. Both now
emit a loud warning naming the affected group / cause.
Two known-degraded states were recorded but never consulted by the read paths, so
a partial result looked authoritative:
- commitSingleOp returns `degraded` ids on an adopt-forward failed-rollback
recovery (the canonical record is durable but its derived-index entry may be
incomplete). persistSingleOp dropped that list on the floor — no health flag,
no read signal. It now records them in a queryable degraded set.
- A non-fatal index-rebuild failure at init (_indexRebuildFailed) was folded into
checkHealth()/validateIndexConsistency() but no read consulted it.
find() and get() now emit ONE loud warning per degraded window (reads still
return — canonical is the source of truth — but the caller is told results may be
partial and to run repairIndex()). Both degraded sources fold into the two health
surfaces, and repairIndex() reconciles from canonical and clears them.
With the unique per-writer temp suffix, a rename ENOENT can no longer mean "a
concurrent idempotent writer already renamed it" — nobody else holds this
writer's temp. It means our just-written temp vanished before the rename, so the
bytes did NOT land. The old code returned success on that ENOENT, acknowledging
a write that persisted nothing; the native provider mmaps these blobs, so a
phantom-acked blob is a silent-loss.
saveBinaryBlob now retries once with a fresh temp (self-healing a transient
external-sweeper/crash-cleanup race), and if the temp vanishes again it throws
loud rather than acknowledging a store-nothing write. Non-ENOENT rename faults
propagate verbatim as before. The unique-temp suffix already eliminated the
concurrent same-key collision that originally motivated the ENOENT shortcut.
The async group-commit flush that persists single-op generation history
swallowed persist failures as a bare warn: writes kept succeeding while their
before-images piled up in memory, never durable and unbounded, with no signal.
The generation store now accounts for every failed flush at one place
(flushPendingSingleOps), tolerates a transient blip (retry with capped
exponential backoff), and after PENDING_FLUSH_FAILURE_THRESHOLD consecutive
failures LATCHES a durability failure and refuses further single-op and transact
writes with a typed, exported PendingFlushDurabilityError — rather than promise a
durability it cannot deliver. Live canonical data is untouched; only the
immutable history is stuck. The latch self-heals: the moment a flush succeeds
(a retry, or an explicit flush()/close()) it lifts and writes resume.
Loud errors, never quiet losses.
clear() removed entities, indexes, system and _cas but left three top-level
trees on disk: _blobs (raw HNSW/LSM segment bytes and the native dkann index),
_id_mapper (the native shared mmap id-mapper), and _column_index (column-store
manifests). A cleared brain therefore re-read stale native state.
Worse, the column store splits its state across two of those trees — manifests
under _column_index/ and their segment bytes under _blobs/_column_index/ — so
removing one without the other stranded a manifest listing segments that no
longer exist, which the hardened segment-load path now (correctly) refuses with
ColumnSegmentLoadError. The three trees now fall together as a set, exactly as
_cas already does. locks/ is deliberately preserved: it is live coordination
state, not data.
ColumnStore silently skipped a manifest-listed segment it could not load: a
corrupt or missing segment dropped every one of its entities out of
filter/rangeQuery/sortTopK with no error, so an inconsistent index read as a
merely short result. It now throws a typed ColumnSegmentLoadError when a listed
segment yields undecodable or no bytes, and lets a genuine storage IO fault
propagate verbatim. Only a field with no manifest at all stays benign (nothing
was ever written for it).
baseStorage.getNoun_internal / getVerb_internal likewise caught every error and
returned null, reporting a present-but-unreadable entity as "not found". They
now return null only for genuine ENOENT-class absence (via isAbsentError) and
rethrow real faults and deserialize errors.
Pattern-B (blind catch) hardening: absence -> null, fault -> loud.
Write/index-spine hardening, first batch of Pass 1. Each fix restores an
invariant the surrounding code already intended; every one has a
fail-before/pass-after test.
- Pattern C, finding 5 (baseStorage): delete now decrements the user-facing
scalar total symmetrically — deleteNounMetadata was decrementing only the
per-type bucket, deleteVerbMetadata neither the bucket nor the scalar, so
getNounCount()/getVerbCount() inflated permanently (the stale scalar wins
pagination via Math.max and is persisted). Invariant now holds:
scalar total === Σ per-type across add/update/delete and reopen.
- Pattern A, finding 3 (graph/lsm/LSMTree): a partial SSTable-load failure no
longer publishes the manifest's full relationship count as healthy. Any
per-SSTable load failure throws after the batch, which resets to honest-empty
and lets the existing size()===0 self-heal rebuild run — size()/isHealthy()
can no longer lie about a partial load.
- Pattern B, finding 6 (hnsw/hnswIndex): deferred flush() no longer clears
dirty nodes whose connections failed to persist — failed nodes stay in the
retry set, and flush() throws HnswFlushError instead of returning a lying
node count. The immediate-mode first-noun saveHNSWSystem is un-swallowed, so
addItem() rejects rather than returning an id for a rootless index.
- Pattern B, finding 11 (part — storage reads): new shared isAbsentError()
helper (utils/errorClassification, ENOENT-only absence) applied to
loadBinaryBlob and readObjectFromPath — a real IO fault (EIO/EACCES/EMFILE)
now propagates loudly instead of masquerading as "absent", which had driven
needless rebuilds / empty reads (loadBinaryBlob feeds the native provider).
Regression: 78 green across the 3 new suites + db-mvcc, generationStore,
temporal-vfs, rollback-trapdoor, restore-nondestructive. Full gate runs before
the Pass-1 release (David-gated). Remaining Pass 1: finding 11 getNoun/getVerb
legs, finding 4 (ColumnStore), finding 8 (pending-flush), finding 10 (degraded),
finding 7 (clear). Pattern A guards (1,2,9) as a follow-up release.
A returned write must be immediately readable by id, metadata filter, vector
search, and graph traversal — no await, delay, or retry between the write
returning and the read. This holds by construction because every projection
(canonical + HNSW + metadata index + graph index) commits inside the write's
transaction; the generation counter is the {seq} a caller can pin. The test
locks that guarantee so index maintenance can never quietly move off the commit
path (which would turn a returned write into a not-yet-queryable one).
First brainy chapter of the write/index-spine hardening program.
When a transaction failed and its rollback ALSO failed to undo a canonical
write (retries exhausted), the old path logged, continued, threw
TransactionRollbackError, and set state='rolled_back' — while the record it
could not undo stayed durable on disk. The caller got an error implying the
write was undone; a read-back showed the record. A failed rollback had no
truthful response (the post-commit response lie).
Give a failed rollback a two-branch honest contract, decided by observation:
both commit paths already hold byte-identical before-images, so after a failed
rollback the store compares current canonical state to them and classifies each
touched record as reconciled, an additive orphan (present when it should be
gone), or a restorative loss (gone/wrong when it should have been restored).
- Adopt-forward: a single-op write whose only damage is a durably-present
orphan — the record the caller asked for — is committed forward (its
generation buffered) and returns success with a loud warning that the derived
index may be incomplete for that id until the next rebuild/repairIndex(). No
error, no double-write; the record is durable and get-able immediately.
- Fail loud + quarantine: a multi-op batch, or ANY restorative loss, throws the
new StoreInconsistentError naming every unreconciled record and its
disposition, and puts the brain into write-quarantine (reads keep working,
writes refused via assertWritable) until repairIndex() reconciles the derived
indexes against canonical and lifts it. The counter is not advanced, and
Transaction.rollback now reports state 'inconsistent' instead of the
'rolled_back' lie when any undo failed.
New: StoreInconsistentError + UnreconciledRecord exported from the root;
repairIndex() forces a rebuild and clears the quarantine. Regression
tests/integration/rollback-trapdoor.test.ts injects index-add-throws +
canonical-delete-undo-fails and pins adopt-forward (durable, get-able, not
quarantined), fail-loud (StoreInconsistentError + quarantine + reads work +
repairIndex lifts it), and the error's record naming.
restore() removed the entire live brain directory and then fs.cp'd the
snapshot in, so any copy failure left the store destroyed with only a partial
copy. The sharpest edge: fs.cp materialized the holes of sparse mmap blob
files, so a snapshot that fits on disk could balloon and ENOSPC mid-copy — the
recovery tool destroying the brain it was asked to recover.
Rewrite restoreFromDirectory as stage → verify → atomic swap:
- Copy the snapshot into a _restore_staging area BEFORE touching live data,
sparse-aware (copyTreeSparse/copyFileSparse skip all-zero 4 MiB chunks, so a
mostly-hole store restores at its true allocated size, not its apparent size).
- On any copy failure (ENOSPC included) remove only the half-written staging
area and throw — the live store is left exactly as it was.
- Only after the copy succeeds, fsync a completion marker naming the staged
entries, then swapStagedRestoreIn(): an idempotent per-entry
rm-old + rename-staged-in (same-filesystem renames that cannot ENOSPC),
removing stale live entries the snapshot lacks.
- completeInterruptedRestore(), wired into init() right after the root dir is
ensured, finishes a crash mid-swap FORWARD (resume a committed swap) or
discards an uncommitted staging area (live still authoritative).
_restore_staging is excluded from snapshots. persist() (hard-link snapshot)
was already safe and is unchanged; no public API change.
Regression (tests/integration/restore-nondestructive.test.ts): a forced copy
failure leaves live data fully intact and cleans staging; a normal restore
round-trips to the snapshot; an interrupted-but-committed restore completes on
reopen; an uncommitted staging area is discarded on open; the sparse copy is
byte-identical with allocated blocks far below apparent size.
A committed transact() reported success while its canonical entity writes were
still only in the OS page cache (tmp+rename, not fsync'd), even though the
generation counter and manifest were fsync'd. A hard kill in that window could
leave the durable counter ahead of the persisted entity bytes — phantom
progress for any generation-based consumer resuming from the counter. Reported
from a downstream migration's crash-lifecycle forensics.
Add an optional transaction durability barrier to GenerationStorage
(beginWriteBarrier / flushWriteBarrier). FileSystemStorage implements it:
writeObjectToPath records each successful canonical write and
deleteObjectFromPath records each delete's parent dir, between begin and flush;
flush fsyncs every recorded write (contents + rename dir entries, via
syncRawObjects) and the parent dir of every delete. commitTransaction opens the
barrier before running the planned operations and flushes it after, BEFORE
persisting the counter and manifest — so the batch's entire canonical footprint
is durable before the generation stamp advances. The barrier is optional: the
generation store calls it through optional chaining, so in-memory and
durable-per-call (cloud object-PUT) adapters treat it as a no-op.
The single-op group-commit path is unchanged (deferred durability is the
Model-B design that avoids a 3-5x per-write fsync regression), but its
durability contract is now documented explicitly on commitSingleOp: transact =
durable on return; single-op = durable at the next flush()/close(), with the
counter buffered alongside the data so a crash loses both together (never a
torn counter-ahead-of-state store).
Regression (tests/integration/transact-durability-barrier.test.ts): the entity
writes fsync in an earlier syncRawObjects batch than the manifest for single-op
and multi-op (add+relate) transactions; a precommit-rejected batch opens no
barrier and advances nothing; MemoryStorage exposes no barrier (optional-chain
no-op).
Transaction.execute() checked its time budget at the top of the operation
loop and threw TransactionTimeoutError from OUTSIDE the per-operation
try/catch, so a mid-flight timeout bypassed rollback entirely — only
per-operation failures rolled back. A bulk transact that crossed its 30s
budget mid-flight left the operations already applied to canonical storage
in place while the generation was never stamped: torn, generation-less
state. The generation-store commit path's abort cleanup explicitly assumes a
throw from execute() already restored the applied operations byte-identically
(it only discards the uncommitted staging directory), so the missing
rollback broke that invariant.
Give execute() a single rollback point: the operation loop is the sole
rollback-guarded region, and any error escaping it — an operation failure OR
a mid-flight timeout — rolls back every applied operation in reverse order,
then surfaces the original error (a rollback failure still supersedes it via
TransactionRollbackError). The per-exit-path rollback that let the timeout
throw slip past is gone; atomicity now holds by construction for every error
type. An aborted transaction leaves generation() unchanged and storage
byte-identical to its pre-transaction state.
Regression (tests/unit/transaction/timeout-rollback.test.ts): a mid-flight
timeout leaves an in-memory canonical store byte-identical with the tx in the
rolled_back terminal state; the operation-failure and rollback-failure paths
through the same single rollback point; and a clean transaction still commits.
transact() promises atomic forward references — add an entity and relate to
it in one batch — but the planner resolved relationship endpoint ints at
PLAN time, before the batch's add operations had applied. Asking the id
mapper about an entity that does not exist yet made the (correctly strict)
native mapper throw in EntityIdMapper.getOrAssign, so
transact([{op:'add', id:X}, {op:'relate', to:X}]) failed on native
deployments; the permissive JS mapper masked the bug and silently leaked an
id assignment whenever a batch was later rejected by a commit precondition
(normal control flow since the conditional-commit CAS landed).
Make endpoint resolution lazy: AddToGraphIndexOperation and
RemoveFromGraphIndexOperation take VerbEndpointInts — an eager
{sourceInt, targetInt} or a thunk evaluated when the operation EXECUTES,
mirroring the constructor's already-lazy generationFn. The four
transact-planner sites (relate, its bidirectional reverse edge, remove's
relationship cascade, unrelate) pass thunks, so resolution happens inside
the commit after same-batch adds have applied; the seven single-operation
sites keep eager objects (their endpoints provably pre-exist — relate
validates both, unrelate/updateRelation/remove read the existing verb).
The remove operation's rollback captures the ints resolved at execute so
its re-add uses the same mappings.
Byproduct fix: a rejected batch no longer pollutes the id mapper — the
regression suite pins this (mapper has no assignment for the phantom
entity after a precondition-rejected forward-ref batch), alongside the
exact reported shape, both-endpoints-in-batch, bidirectional,
add+relate+remove in one batch, and the split-transact control
(tests/integration/transact-forward-ref-graph.test.ts).
The temporal model had a hole exactly where files were concerned: every
entity write is an immutable generation with before-images, but VFS content
BYTES lived under an eager refCount GC left over from the pre-8.0 design —
unlink could physically destroy bytes that in-window history still
referenced, and overwrite never released the old hash at all (an unbounded
silent leak whose accidental byproduct was the only thing "preserving"
history). Reading the past could therefore return a stale field, a dangling
hash, or nothing, depending on luck.
Fix: blob reclamation becomes a HISTORY decision instead of a LIVENESS
decision. Each blob's metadata now carries historyRefCount alongside the
live refCount:
- The commit seam counts one history reference per persisted before-image
record carrying a content hash (commitTransaction staging and the
group-commit flush), recorded BEFORE the record-set persists and carried
in the generation delta (blobHashes — always present on new deltas, so
compaction only falls back to reading records for pre-contract
generations). An aborted transaction compensates best-effort.
- unlink/rmdir/overwrite drop ONLY the live reference (BlobStorage.delete →
release; overwrite finally releases the superseded hash — cancelling the
dedup increment on same-content rewrites and closing the leak), and only
AFTER the canonical mutation commits, so a failed delete can never leave a
live file whose bytes compaction might reclaim.
- History compaction is the ONE reclamation point: after deleting a
generation's record-set it releases that set's references and physically
reclaims any hash at zero live AND zero history references. Pins are
exempt automatically. Crash ordering is over-count-only in every path
(record before persist, release after delete), so a crash can leak until
the scrub recounts but can never reclaim bytes a retained generation
needs. scrubBlobHistoryRefCounts() restores exactness; existing stores get
a one-time marker-gated backfill on open, failing into leak-safe mode
(reclamation disabled) rather than guessing.
On top of the protected history, the temporal API the generational model
always implied:
- vfs.readFile(path, { asOf }) — the exact bytes as of a generation or Date,
materialized from the history (pinned view released so compaction is
never blocked by a read).
- vfs.history(path) — FileVersion[] ascending ({ generation, timestamp,
hash, size, mimeType? }), the newest entry being the live state.
- Overwrites now refresh the file entity's data/embedding text — semantic
search and the data field previously served the FIRST version's text
forever (the stale-field defect a consumer's incident recovery depended
on by luck).
Integration suite (temporal-vfs.test.ts): per-version exact reads +
history listing, leak-fix + history protection on overwrite, rm keeps bytes
readable, compaction reclaims past-window bytes and preserves in-window
(including the cross-file dedup case where an old file's history and a
newer file's removal share one hash), data freshness, and scrub exactness.
Subscribe once and receive one post-commit event per affected record for
EVERY canonical write, regardless of origin: direct calls, batch methods,
transact(), imports, and Virtual Filesystem writes all funnel through the
same commit points the feed is emitted from. This is the authoritative
in-process signal for live UIs, cache invalidation, and realtime sync layers
that forward it over their own transports.
Architecture — the post-commit dual of the ifRev precommit hook: mutation
methods hand lightweight event descriptors to the commit seam
(persistSingleOp / transact's plan), which stamps the committed
{generation, timestamp}, enriches entity deletes with the record's LAST
committed state from the commit's own before-images (free — Model B reads
them anyway; removeMany's id-only deletes gain full payloads this way), and
emits only after the commit succeeds — a losing CAS or rejected batch never
announces anything. Dispatch is a microtask FIFO after the mutex releases:
commit-ordered, a slow listener never delays a write, a throwing listener is
logged and isolated, and with no subscribers the write path constructs no
events at all. Single-point emission was chosen over per-method hooks
because the aggregation-hook pattern demonstrably drifted (relation ops and
removeMany were silently missing from it).
Coverage: add/update/remove (+ cascade unrelate per deleted relationship),
relate (both edges when bidirectional)/unrelate/updateRelation, per-item
events for addMany/updateMany/relateMany/removeMany, per-item events sharing
one generation for transact(), and transitively imports + VFS. clear() and
restore() — wholesale raw-state operations outside the per-record commit
path — emit a single store-level event meaning "refetch everything".
brain.close() drops all listeners.
New module src/events/changeFeed.ts (BrainyChangeEvent + ChangeFeed,
exported from the package root); guide docs/guides/reacting-to-changes.md.
Integration suite pins the contract: per-op payload fidelity, delete
last-state payloads, batch per-item emission, one-generation transact
batches, VFS-origin events, CAS-loser silence, ordering, listener isolation,
unsubscribe, and store-level events.
Sweeping the vestigial hnsw sharding machinery surfaced two real defects on
the counts-recovery path (the code that rebuilds totalNounCount/totalVerbCount
when counts.json is lost or corrupted — container restarts, partial copies):
- initializeCountsFromDisk counted files in entities/*/hnsw/ — directories the
8.0 write path never populates — so an established store recovered to ZERO
counts on real data: wrong getNounCount()/stats, a "New installation"-style
boot log, and a mis-sized rebuild-strategy decision at open. The scan now
walks the canonical entities/<kind>/<shard>/<id>/ tree (one entity per id
directory), with the same sampled type-distribution estimate as before.
- The sampler called getNounMetadata — whose ensureInitialized() re-enters
init() — from INSIDE init(), a latent deadlock reachable the moment the scan
found anything to sample. The sampled metadata files are now read directly
with fs (gz-transparent), no guarded accessors inside init.
The dead machinery itself (verified zero callers by reachability analysis
before deletion): the nine 7.x hnsw-layout entity/edge CRUD methods, the
sharding-depth prober + its depth-migration engine (unreachable — the probe
always returned null on 8.0 stores), an orphaned streaming verb paginator,
their path/scan helpers, and the now-unused type aliases and fields. The init
block keeps the truthful new-vs-established boot log introduced in 8.0.13,
now decided directly from the canonical layout. Net -1,138 lines; no public
API touched.
Adds a regression test: delete counts.json from a populated store, reopen,
counts recover exactly from the canonical tree.
The fast-follow to the ifRev CAS fix: the sweep for the same check-then-act
class found two more instances, both now closed with the same discipline —
the decision runs at the serialization point that guards the apply.
add({ifAbsent}) / add({upsert}): the absence check ran before the commit
mutex, so N concurrent same-id creates could all pass it and all write — the
second silently overwriting the first, violating ifAbsent's "return the
existing id WITHOUT writing" contract and upsert's merge-never-clobber
contract. The insert leg now carries a must-be-absent commit precondition
(the same conditional-commit primitive ifRev uses), verified under the commit
mutex against the authoritative before-image. A losing caller takes its
documented resolution instead of overwriting: ifAbsent returns the existing
id with zero writes; upsert merges into the now-existing entity via update()
(shared param mapping in upsertMergeParams so the planning-time branch and
the conflict retry can never drift), with a bounded retry if a concurrent
delete lands between the conflict and the merge. The planning-time checks
remain as fast-fails. transact()'s batch-level ifAbsent/upsert keep
planning-time semantics (documented, converges to a valid entity).
BlobStorage: write()'s dedup decision (exists → increment refCount, absent →
create at 1) and delete()'s decrement-then-remove were unserialized
read-modify-writes over blob-meta:<hash>. Concurrent writes of identical
content could lose references, so a later delete removed bytes another file
still referenced (data loss), or leaked unreferenced blobs. All
reference-count-bearing mutations now serialize through a per-hash
InMemoryMutex — distinct content never contends; incrementRefCount/
decrementRefCount are documented lock-assumed internals.
Both fixes are complete by construction in-process: storage enforces
single-writer-per-directory, so the process is the whole concurrency domain.
Tests (tests/integration/ifabsent-upsert-blob-concurrency.test.ts): 8-way
ifAbsent storm advances the store by exactly one generation with _rev 1;
8-way upsert storm on an absent id yields one create + seven merges
(_rev === 8); sequential ifAbsent/upsert semantics pinned unchanged; N
identical concurrent blob writes → refCount === N; the blob survives until
the true last reference drops; interleaved write/delete storm never loses a
landed reference.
2026-07-08 15:13:26 -07:00
330 changed files with 55346 additions and 5638 deletions
@soulcraft/brainy (v7.17.0) is a Universal Knowledge Protocol -- a Triple Intelligence database combining vector search, graph traversal, and metadata filtering in a single library. Published to npm as a public MIT-licensed package.
@soulcraftlabs/brainy (v7.17.0) is a Universal Knowledge Protocol -- a Triple Intelligence database combining vector search, graph traversal, and metadata filtering in a single library. Published to npm as a public MIT-licensed package.
# The release script bumps package.json's version before it tags, so
# this tag's checkout already carries the version being published —
# nothing here re-derives it from the tag name.
PUBLISH_OK=true
if ! npm publish --tag "$NPM_TAG" --userconfig "$TMPRC"; then
PUBLISH_OK=false
fi
# Readback verify is the source of truth, run regardless of the publish
# exit code: a benign duplicate publish (a prior run, or a mirror, already
# landed this exact version) reports failure even though the registry
# already holds the right content.
LANDED_VERSION="$(npm view "@soulcraftlabs/brainy@${VERSION}" version --userconfig "$TMPRC" 2>/dev/null || echo "")"
rm -f "$TMPRC"
if [ "$LANDED_VERSION" != "$VERSION" ]; then
echo "::error::Readback verify FAILED — The Source registry reports version '${LANDED_VERSION:-<none>}', expected '${VERSION}'. This is a genuine publish failure, not a benign duplicate."
exit 1
fi
if [ "$PUBLISH_OK" = true ]; then
echo "Published and verified @soulcraftlabs/brainy@${VERSION} on The Source registry."
else
echo "::warning::npm publish reported failure, but readback confirms @soulcraftlabs/brainy@${VERSION} is already live on The Source (a prior run or mirror landed it) — treating this run as successful, since the registry content is correct. Any OTHER failure mode would have failed the readback check above instead."
All notable changes to this project will be documented in this file. See [standard-version](https://github.com/conventional-changelog/standard-version) for commit guidelines.
- docs(releases): 10.4.1 and 10.4.2 consumer notes; 10.4.2 is the last MIT release under this name, Open Brainy continues at @soulcraftlabs/brainy (a082e0ef)
- fix(vectors): a zero-norm vector is not a vector, canonical side included, plus the sanctioned unvector door (0de76659)
- fix(hnsw): skip unvectored rows on rebuild; refuse empty vectors in the index (8fc553b1)
- fix(storage): derive the canonical count ledger from identity records, stamp the derivation rule, and mark legacy-derived ledgers suspect at load (fd6b4ce4)
- docs(concepts): the serving law — a failure is graded by whether an answer could be wrong, never by the cost of the fix; reads refuse per family (2914e0eb)
- docs(releases): the 10.4.0 entry catches up to the late trains — repair routing, the vector ledger and open-gate leg, the loud config guard, the JSON-safe crossing (834149ed)
- test(readiness): the report helper's clock freezes — two independently-built reports compared across a millisecond tick made the plant lane red (39b916a3)
- feat(repair): a heal:'repair' verdict routes to the provider's own incremental repair() (553e0d97)
- fix(storage): an unknown nested storage config can never silently land on the shared default root (ddd5e719)
- docs(release): the 10.4.0 entry, the index-health concept doc, and the API surfaces — written from the tree, not the plan (8cced871)
- fix(plugins): the silent-degrade doors close — a broken accelerator install can never read as absent (b9ba50fb)
- feat(recovery): the catchup verdict is consumed; verb rows go live; the metadata rebuild goes online (18f172e0)
- feat(health): the gate reads the named report — reads refuse loudly, never rebuild; open serves before it returns; the ceremony door (f8f64780)
- docs(releases): the 10.1.0 consumer entry — bounded recovery, restore founding, the two write-path cures (7d3c8696)
- fix(restore): a restore is an unclean event — the swap runs quiesced and the snapshot's durability stamps never survive it (9ca80667)
- feat(recovery): the fold-checkpoint bound — crash folds (checkpoint, head], never the whole log twice (ff43de1a)
- fix(log): pad-frame construction is total; the at-ack sync-failure compensation splits by phase — a production adoption's two write-path defects, cured at their roots (cbe34d11)
- feat(query): the sparse-store cut — where on a never-carried field serves operator truth, never a refusal (7b67db4d)
- fix(adoption): the baseline backfill cures hydration-law drift — existing brains reach the crash-safe default with zero operator steps (25f0dd96)
- fix(adoption): the reserved-root mint exemption — int 0 is legitimate for exactly one id (2abe8b38)
- fix(recovery): walks are healers — the typed/tolerant boundary redrawn where block-layer fault injection proved it belonged (0e3facf4)
- feat(log): log authority is the fleet default — adopt-at-open, oracle-gated; plus the power-cut throw-site cures and the loud torn-record contract (214c98b4)
- fix(durability): three block-layer power-loss findings from the first fault-injection box run — all cured, matrix 15/15 (67c606be)
- docs: RELEASES.md frames the release as 10.0.0 — honest major (log format v2 forward-only); comment wording cleanup (d1698fa5)
- fix(persistence): the idle flush trigger debounces under load — deferred to the floor, never dropped, never a flush-per-gap amplifier (a50726e6)
- feat(reprojection): the one doors-open machinery — budget-capped, yielding, foreground-preempted, atomic-swap; poison records quarantine typed (d1651f98)
- feat(embedding): deferred-embed markers become log records — the sidecar recovery path is deleted (b47787bb)
- feat(conformance): the golden-log fold oracle — encoder bytes and fold semantics pinned by content hash (c95bea88)
- feat(engine): the wiring wave — stamps ride every flush, provider generations, waitForIndexed, adopt-backfill, match-all serves (b53e6e89)
- feat(index): watermark stamps on every TS projection — adopt/catchup/rescan verdicts at load, stamp-after-data (b35d87a7)
- feat(log): v2 is the LIVE write format — envelope records with minted ints, genesis, sector seals; v1 readable forever (26c60251)
- docs: RELEASES.md — the unreleased write-path and lifecycle entry (consumer-facing draft; version set at cut) (73eb88d4)
- feat(temporal): as-of semantic recall joins the release contract — past vectors byte-exact, pinned (f7ca0d26)
- fix(log): acked writes survive power loss; rejected writes never silently commit — the kill-matrix goes 11/11 with zero .fails debt (13022c51)
- feat(plugin): every provider write surface carries the real committed generation (2d532684)
- feat(log): fact-log format v2 codec — record envelope, type registry, genesis, sector seals; fault-injection shim (34841074)
- feat(log): the guarded log-authority core — group-commit durable-at-ack, the per-brain switch, the verification oracle (65953097)
- docs: Path Registry rows DP6/DP8/MT5 flip to contracted+pinned — the deferred-embedding and atomic-update train landed with cited tests (9fda6d95)
- feat(embedding): MT5 — deferred embedding with durable markers; write acks never wait on a neural net (287384cf)
- fix(index): the flicker window dies — atomic in-place vector update; lazy open honors every provider's not-ready report; the Path Registry twin table (ebe06cdf)
- feat(persistence): the engine owns its flush cadence — callers never call flush() in hot paths again (3236a01b)
- fix(aggregation): the lifecycle cluster — flush stamps, behind-stamp catches up incrementally, the native rebuild finally gets invoked, deletes are never silently skipped (1dc861d2)
- perf(sort): ordered reads never do per-row storage round-trips — the 199-317s production scan class dies structurally (607b6b56)
- chore: the home registry is The Source, never 'the forge' — sweep the misnomer out of the release rail, workflows, and release notes (Forge is a different product; the stored CI secret keeps its historical name) (09352c2b)
- ci: tags stop triggering the CI matrix (redundant re-run of already-tested commits starved every release's publish run on the sequential runner) + release.sh forge poll window 20→50 min (c6c6ea6b)
- test: version-coupling pins go major-agnostic — the 8.x literals broke at the 9.0.0 bump while the coupling law itself behaved correctly (8a6807e8)
- docs: 9.0 namespace-migration guide — the simple story + the mechanical sweep checklist, published for humans and tooling alike (61ab9db2)
- fix(release): storefront leg republishes CI's exact forge artifact — byte-identity by construction, verified by cross-registry shasum before the ceremony reports success (d89df2ed)
- docs: v9.0.0 release notes — the field-addressing law migration ledger; retitle the shipped 8.11.0 canonical-enumeration entry (header went stale at its cut) (55a7512c)
- feat(namespace): merge the field-addressing law train — no special names, system.* scalars, nested-bag storage, epoch-3 index keys (19b477ae)
- feat(namespace): NO SPECIAL NAMES + storage fidelity — the ruled completion of the field-addressing law (24bf6cdb)
- feat(namespace): write-door forgery refusal (user metadata keys may never start 'system.') + refusal messages name both spellings in every branch (the non-colliding case marks system.<f> honestly as NOT valid) — cross-engine message pin alignment (48a6130a)
- feat(namespace): conformance green 19/19 — data-aware did-you-mean on unindexed bare addresses, ordering contract on the column top-K path (never drop, nulls last, ties by id), shape-complete addressed reads (entity views AND raw storage shapes, shadow-proof both scopes), per-key source matching for dotted addresses; refusal classes unified under UnresolvableFieldError (8e962dab)
- feat(namespace): aggregation reads under the law + epoch 3 (the key-split rebuild) + THE ARMING COMMIT — the capability constant, the law module, and the typed refusals export from the package root; both engines' conformance suites light on this signal (7492b6cb)
- feat(namespace): egress guard + validation speak the law — whereMatcher's resolver reads system.* from the record and bare names from the metadata bag only (the bare-system switch is dead); validateFindParams refuses cursor/includeRelations/writeOnly typed (accepted-and-ignored dies as a class), validates order, and parses every orderBy address (c2fb28a2)
- fix(namespace): noun-record updates preserve legacy inline HNSW adjacency — the placeholder-adjacency write stamped out pre-codec records' stored connections (crash-window unreachability); codec-era records were never at risk (empty field is the blob marker); pin covers the legacy shape (4679c894)
- feat(namespace): find's own filter builders speak the frozen keys — params.type/subtype/service become system.* index keys at every construction site (three pipelines + the canonical buildMetadataFilter); the where.type→noun alias is dead (bare 'type' belongs to the user now) (7a28a946)
- feat(namespace): the index speaks the frozen keys — record-frame scalars index under literal 'system.<field>' (legacy 'noun' spelling folds into system.type; plumbing never indexed from a record frame), user fields stay bare in every shape; filter + sorted paths route every address through parseFieldAddress; storage fallbacks read the addressed side of the record (11c724bc)
- docs(namespace): the d.ts JSDoc wave — the sealed field-addressing law on the full find + aggregation surface, present-tense, with the refusal semantics and migration note inline (comment-only; verified zero code lines changed) (fcb24ab6)
- test(namespace): unit pins for the pure law — the ruled maps verbatim (incl. the relation mirror, unpinnable via public API), plumbing refusals both kinds, did-you-mean text (5502abcd)
- fix(namespace): the JS sorted fallback honors the ruled ordering contract — nulls last in BOTH directions (was nulls-first on desc) + deterministic id-ascending tie-break (56deb2e8)
- test(namespace)+docs: the cross-engine conformance suite (self-arming — skips until the resolver exports land) + the public field-addressing docs page; sidebar order deconflicted to 7 (d8d0b55f)
- feat(namespace): the one field-addressing law as a single source of truth — parseFieldAddress + the ruled ten-scalar system maps + plumbing invisibility + refusal builders (module only; query surfaces wire in next) (8f9a9989)
- docs: port the 8.10.3 backport-release changelog entry to main (f6b14d21)
- docs: port the 8.10.2 backport-release changelog entry to main — release branches carry the version bump, main carries the durable record (0b059ac5)
- fix: user metadata named 'level' is a real field everywhere — the engine-internal node layer no longer shadows it in sort/filter/aggregation, and the indexing views stop stamping a phantom 0 into its column; index epoch 2 rebuilds existing brains at first open (1a09be06)
- fix: metadata-only update() never rewrites the noun record — the unconditional whole-vector save turned per-entity stat touches into full rewrites+fsync, amplifying read-heavy sweeps into disk saturation on a production deployment (cb717be2)
- fix(release): double the forge-publish poll budget — the runner executes jobs sequentially and the publish run queues behind the ci matrix (64049631)
- Merge branch 'release/8.11.0' (1865f60a)
- Merge branch 'release/8.10.1' (fc9f0d72)
- chore: the forge is the address — retire the archived mirror from every live surface (415e824a)
- docs: the last two archived-host links point home (91ef1c8b)
- feat: includeHidden — export carries every visibility tier for migration-grade canon completeness (63c1eeb9)
- feat(release): the forge publish leg moves to CI on the tag push; the laptop verifies by readback and keeps the abort-before-storefront guard (3e4a17dc)
- feat: canonical enumeration mode for export — storage-walked, canon-complete, with an index-drift report (4d196af4)
- docs: dedupe the 8.10.2 release-notes entry the cherry doubled onto the branch (8c956608)
- fix: user metadata named 'level' is a real field everywhere — the engine-internal node layer no longer shadows it in sort/filter/aggregation, and the indexing views stop stamping a phantom 0 into its column; index epoch 2 rebuilds existing brains at first open (958a0859)
- fix: metadata-only update() never rewrites the noun record — the unconditional whole-vector save turned per-entity stat touches into full rewrites+fsync, amplifying read-heavy sweeps into disk saturation on a production deployment (5b65eb82)
- refactor: remove the orphaned transaction-result type left behind by the dead-path removal (edf123a5)
- fix: warm() metadata surface routes through the active provider (warm hook added to the metadata contract); add maintenanceDebt() observability surface (5b2cbf74)
- fix: transaction timeouts are a typed no-hot-retry contract; engine-side non-retry pinned; dead transaction path removed (003e2a74)
- chore: the forge is the address — retire the archived mirror from every live surface (22702b81)
**Brainy's current open actions:** None. MIT open-source — no platform-specific actions.
**Current version:** `@soulcraft/brainy@7.31.5` (latest published; 8.0.0 release candidate on `feat/8.0-u64-ids`)
**Current version:** run `npm view @soulcraftlabs/brainy version --registry https://source.soulcraft.com/api/packages/soulcraftlabs/npm/` (never trust a hardcoded number here — this line went stale for months); consumer-facing changes tracked in `RELEASES.md`
---
## Project Overview
Brainy is a Universal Knowledge Protocol -- a Triple Intelligence database that combines vector similarity search, graph traversal, and metadata filtering into a single TypeScript library. Published as `@soulcraft/brainy` on npm under the MIT license.
Brainy is a Universal Knowledge Protocol -- a Triple Intelligence database that combines vector similarity search, graph traversal, and metadata filtering into a single TypeScript library. Published as `@soulcraftlabs/brainy` on The Source (source.soulcraft.com registry) under the MIT license.
## Getting Started
@ -91,7 +91,7 @@ test: add/update tests (patch version bump)
## Docs Pipeline — soulcraft.com/docs
Docs in `docs/**/*.md` are published with the npm package (included in `files`) and synced to soulcraft.com/docs on every portal deploy. Frontmatter controls what appears publicly.
Docs in `docs/**/*.md` are published with the npm package (included in `files`) and go live on soulcraft.com/docs via the docs ingest API: the release script's `scripts/push-docs.js` step POSTs every public doc to `https://soulcraft.com/api/docs/ingest` (auth: `DOCS_INGEST_SECRET` in the environment). No separate deploy step is involved (the old deploy-to-publish flow was retired in a platform change, 2026-08). Frontmatter controls what appears publicly.
The script: verifies clean git state, builds, tests, bumps version, updates CHANGELOG.md, commits, tags, pushes, publishes to npm, and creates a GitHub release.
After a successful release, remind the user:
> "Published. Deploy portal to pick up the new docs → go to the portal project and deploy."
> "Published. Docs are live on soulcraft.com/docs (pushed via the ingest API during the release) — spot-check a changed page with curl."
Do NOT deploy portal from here. Portal is always deployed separately from within the portal project.
There is no separate deploy step anymore. If the docs push failed (the script warns loudly), re-run `node scripts/push-docs.js` with `DOCS_INGEST_SECRET` set.
<ahref="https://source.soulcraft.com/soulcraftlabs/-/packages/npm/brainy"><imgsrc="https://img.shields.io/badge/package-The%20Source-2c3e50.svg"alt="Package on The Source"></a>
<ahref="#from-laptop-to-hundreds-of-millions">Scale with Cor</a> ·
<ahref="#documentation">Docs</a>
<ahref="#when-you-outgrow-brainy">Scale with Cor</a> ·
<ahref="#documentation">Docs</a> ·
<ahref="#support--community">Support</a>
</p>
---
**Open Brainy** is the MIT engine — the open API, client library, types, and protocol; an openly specified canonical on-disk format; and this TypeScript reference engine, scoped as a single-node engine for stores up to roughly one million rows. `@soulcraft/brainy` 10.4.2 was the last release under the old package name — the name passes to the native engine, **Brainy**, at 11.0.0: the same API over the same open format at production scale, and it requires a license.
Built because we were tired of stitching a vector store to a graph database to a document store — and spending weeks on plumbing before writing a line of business logic. Brainy indexes every fact **three ways at once** and lets one call query them together:
| You write | Brainy indexes it as | You query it with |
@ -44,12 +47,14 @@ It runs **inside your process** — no server, no Docker, nothing to operate —
## Quick start
```bash
bun add @soulcraft/brainy # Bun ≥ 1.1 — recommended
npm install @soulcraft/brainy # Node.js ≥ 22
bun add @soulcraftlabs/brainy # Bun ≥ 1.1 — recommended
npm install @soulcraftlabs/brainy # Node.js ≥ 22
```
> **Registry**: add `@soulcraftlabs:registry=https://source.soulcraft.com/api/packages/soulcraftlabs/npm/` to your `.npmrc` (anonymous read).
```javascript
import { Brainy, NounType, VerbType } from '@soulcraft/brainy'
import { Brainy, NounType, VerbType } from '@soulcraftlabs/brainy'
const brain = new Brainy() // in-memory; one line swaps to disk
Brainy's TypeScript engines take you a long way. When you outgrow them, add the native engine — **the API doesn't change**:
Brainy's pure-TypeScript engines carry real workloads a long way on their own — see the measured, per-operation numbers (not marketing figures) in **[docs/performance-envelopes.md](docs/performance-envelopes.md)** for what to expect, unaccelerated, on plain filesystem storage.
When a deployment needs native-scale vector/graph performance — memory-mapped indexes that don't need your dataset in RAM, billion-scale ambitions — add the native engine. **The API doesn't change:**
```bash
npm install @soulcraft/cor
@ -187,13 +194,14 @@ await brain.init() // @soulcraft/cor detected — same code, native engines un
Installing the package is the opt-in: if `@soulcraft/cor` is present, it loads and announces itself in the init log; if it's present but broken, `init()`**throws** — an installed accelerator never silently vanishes behind the JS engines. Opt out with `plugins: []`, or pin exactly what loads with `plugins: ['@soulcraft/cor']`. [`@soulcraft/cor`](https://www.npmjs.com/package/@soulcraft/cor) (Brainy 8.x ↔ Cor 3.x, version-matched) registers Rust implementations behind every provider seam: SIMD distance kernels, memory-mapped storage, a disk-native vector index that doesn't need your dataset in RAM, durable LSM field/graph indexes that serve cold opens instantly, and native aggregation. Recall@10 measured **0.99 / 0.96 / 0.96 at 1M / 10M / 100M vectors** in Cor's release gate.
Open core, commercial accelerator: Brainy is MIT and complete on its own; Cor is licensed and funds both.
Open core, commercial accelerator: Brainy is MIT and complete on its own — Cor is more headroom for when you need it, not capability held back to sell you later. Licensing and support: **cor@soulcraft.com**.
## Performance
- Per-operation p50/p95 at 1k and 10k entities, pure-JS floor, measured and re-run every release that touches a measured path: **[docs/performance-envelopes.md](docs/performance-envelopes.md)**.
- JS distance kernels: **~6× faster cosine, ~1.4× euclidean** than 7.x (measured: [`tests/benchmarks/distance-microbench.mjs`](tests/benchmarks/distance-microbench.mjs), 384-dim, median of 41).
- Whole-graph reads are single **O(N + E)** cursor walks — a consumer-measured 19k-edge export dropped from ~27 s of per-node calls to one scan.
- Full numbers and capacity planning: **[docs/PERFORMANCE.md](docs/PERFORMANCE.md)** · **[docs/SCALING.md](docs/SCALING.md)**
- Capacity planning and architecture: **[docs/PERFORMANCE.md](docs/PERFORMANCE.md)** · **[docs/SCALING.md](docs/SCALING.md)**
## Use cases
@ -212,6 +220,10 @@ Open core, commercial accelerator: Brainy is MIT and complete on its own; Cor is
**Bun ≥ 1.1** (recommended) or **Node.js ≥ 22**. Brainy 8.x is server-only; the 7.x line remains on npm for browser use.
Brainy has a plugin system that allows third-party packages to replace internal subsystems with custom implementations. This is how `@soulcraft/cor` provides optional native acceleration, and it's the same system available to any developer.
@ -46,7 +46,7 @@ If no plugin provides a given key, brainy uses its built-in JavaScript implement
### 1. Implement the `BrainyPlugin` interface
```typescript
import type { BrainyPlugin, BrainyPluginContext } from '@soulcraft/brainy/plugin'
import type { BrainyPlugin, BrainyPluginContext } from '@soulcraftlabs/brainy/plugin'
const myPlugin: BrainyPlugin = {
name: 'my-brainy-plugin', // Must be unique (typically your npm package name)
@ -90,7 +90,7 @@ await brain.init()
**Programmatic registration:** For plugins not installed as npm packages, use `brain.use()`:
@ -246,10 +272,10 @@ When provided by an optional native acceleration plugin (such as `@soulcraft/cor
#### `cache`
**Type:** `UnifiedCache`
Replaces the global `UnifiedCache` singleton used for VFS path resolution, semantic caching, and vector index caching. Must implement the `UnifiedCache` interface (available from `@soulcraft/brainy/internals`).
Replaces the global `UnifiedCache` singleton used for VFS path resolution, semantic caching, and vector index caching. Must implement the `UnifiedCache` interface (available from `@soulcraftlabs/brainy/internals`).
```typescript
import type { UnifiedCache } from '@soulcraft/brainy/internals'
import type { UnifiedCache } from '@soulcraftlabs/brainy/internals'
context.registerProvider('cache', myNativeCache)
```
@ -299,8 +325,8 @@ Plugins can register custom storage backends that users reference by name.
### Implementing a Storage Adapter
```typescript
import type { StorageAdapterFactory } from '@soulcraft/brainy/plugin'
import type { StorageAdapter } from '@soulcraft/brainy'
import type { StorageAdapterFactory } from '@soulcraftlabs/brainy/plugin'
import type { StorageAdapter } from '@soulcraftlabs/brainy'
class MyStorageAdapter implements StorageAdapter {
async init(): Promise<void> { /* ... */ }
@ -334,9 +360,9 @@ Brainy provides three entry points for plugin developers:
> **Stale as of 10.4 — "Mode 2: Lazy Loading on First Query" below is RETIRED.**
> `disableAutoRebuild` no longer defers index construction to a first query;
> `brain.init()` now runs every needed rebuild to completion before it
> returns, unconditionally. A read against a not-serving provider throws a
> typed `*NotReadyError` instead of rebuilding mid-query. See
> `docs/concepts/index-health.md` for the current contract; this document's
> line-number references to `src/brainy.ts` also predate the file's current
> size and are unreliable. Left as historical background on the rebuild
> mechanics, not as a current API description.
This document explains how Brainy's four indexes (MetadataIndex, vector index, GraphAdjacencyIndex, DeletedItemsIndex) initialize and rebuild from persisted storage.
- **Result: Model data across virtually any industry**
Every piece of information can be represented as entities (nouns) connected by relationships (verbs) carrying properties (metadata). The standardized type system from `@soulcraft/brainy` (`NounType`, `VerbType`) gives those nouns and verbs a stable, shared name.
Every piece of information can be represented as entities (nouns) connected by relationships (verbs) carrying properties (metadata). The standardized type system from `@soulcraftlabs/brainy` (`NounType`, `VerbType`) gives those nouns and verbs a stable, shared name.
## The Power of Standardization: Universal Interoperability
title: Field addressing: your fields and system fields
slug: concepts/field-addressing
public: true
category: concepts
template: concept
order: 7
description: The one rule for every query-surface field name — a bare name always means your metadata, system.<field> reaches the ten engine scalars explicitly, and anything else refuses by name.
next:
- guides/namespace-migration
- concepts/consistency-model
---
# Field addressing: your fields and system fields
Every query surface in Brainy — `find()`'s `where`, `orderBy`, aggregation
`groupBy`, and aggregation `source.where` — resolves field names by one rule,
with no exceptions:
> **A bare field name always means your metadata. `system.<field>` reaches an
> engine scalar, and only when you spell it explicitly.**
```typescript
await brain.find({ orderBy: 'level' }) // reads entity.metadata.level — YOUR field
description: Every committed write also appends a self-verifying "fact" — an after-image commit record — to an append-only log. What facts are, the crash-safety model, the scanFacts() streaming surface, family stamps, and how index providers consume the log for sequential heals.
next:
- concepts/consistency-model
- guides/snapshots-and-time-travel
---
# The Generation Fact Log
Since 8.4.0, every committed generation also appends a **fact** — a compact record of what each
touched entity or relationship *became* — to an append-only, checksummed log under
`_generations/facts/`. Where the generational history answers *"what did things look like
before?"* (before-images, powering `asOf()` and rollback), the fact log answers *"what happened,
in order?"* — one sequential, self-verifying stream of the store's present being written.
Nothing about querying changes. The fact log exists for three consumers:
1. **Index heals and rebuilds** — one sequential read in commit order replaces a per-entity
directory walk over millions of files.
2. **Incremental catch-up** — a derived index that knows which generation it reflects reads *just
the gap*, instead of rebuilding from scratch.
3. **Replay and audit tooling** — anything that wants the store's committed timeline as a stream.
## What a fact is
One fact per committed generation:
- **`generation`** and **`timestamp`** — which commit, when.
- **`ops`** — every write in that commit: `{ kind: 'noun' | 'verb', id, record }` where `record`
holds the entity's full after-image (both stored legs), or **`null` for a tombstone** — a
removal carries no body, by design.
- **`meta`** — the transaction metadata `transact()` was submitted with, when present.
- **`blobHashes`** — content-blob references, for exact reclamation accounting.
Facts accumulate **from the first write after upgrading** — pre-existing history is not
retroactively converted, and consumers fall back to the enumeration walk when no log exists.
## Crash safety, in one paragraph
Facts are appended and fsynced **inside the same durability window as the commit itself**, before
the commit point — so after a crash, the log can only ever be *ahead* of committed truth, never
behind it with a hole. On open, the store reconciles the log back to the committed watermark:
torn tails are detected by per-record checksums and cut; whole records beyond the watermark are
truncated. The invariant every reader can rely on: **an absent generation was never committed; a
present fact was.** `transact()` facts are durable the moment `transact()` returns; single-op
facts share the same group-commit flush as the rest of their generation, so a hard kill loses the
fact and the generation *together* — never a torn state.
description: How Brainy knows whether a derived index can be trusted — exact accounting instead of sampling, the named health report, degraded-but-serving vs. not-ready, and what repairIndex() checks, heals, and rebuilds.
next:
- concepts/generation-fact-log
- guides/inspection
---
# Index Health
Brainy keeps one **canonical** copy of every entity and relationship, and three
**derived** indexes built from it — vector, metadata, and graph — so `find()` can
answer semantically, by filter, and by traversal without re-deriving the answer from
scratch on every query. A derived index is a cache with a serving structure: it can
be present but stale, present but only partially loaded, or fully out of sync with
canonical after a crash. This page is about how Brainy decides whether to trust one,
what it does when it can't, and how you reconcile the two.
## Exact accounting instead of sampling
Older health checks worked by inference: does `size()` return something greater
than zero, does a spot-check on one known item come back correct. Both are proxies.
A cold index can report a nonzero count while its actual serving structure never
loaded, and a spot-check only proves the one item it happened to ask about.
Every derived-index provider may now expose a named, synchronous, O(1)
`healthReport()` — composed from the provider's own **exact ledgers** (real counters
it already maintains on the write path), never a sample or a walk. This is the one
signal Brainy's read gate consults. A provider that doesn't yet expose one falls
back to an honest `isReady()` boolean, and finally to a size heuristic for engines
with neither — but wherever a `healthReport()` exists, it wins.
Underneath, storage itself keeps an analogous **canonical count ledger**: a
`counted` scalar (the user-facing total — what `getNounCount()` / `getVerbCount()`
return) and an `all` scalar (every tier, including internal records a derived
index's own coverage math needs to compare against). This is the real denominator
a provider's `healthReport()` measures itself by, rather than a total that can only
ever ratchet upward. See [What `suspect` counts mean](#what-suspect-counts-mean)
below for the one case that ledger can't stay exact through on its own.
## The named report
A `HealthReport` carries, per provider (`'vector'` / `'graph'` / `'metadata'`):
- **`healthy`** — `true` iff every *verified* invariant holds. An invariant whose
family has no ledger yet is `unledgered`, never counted either way — unknown,
not passing.
- **`serving`** — can this provider answer a query right now. A failing invariant
graded `heal: 'repair'` or `heal: 'none'` still leaves `serving: true` — this is
**degraded-but-serving**: something is off (say, a stale rollup on an
`employee` record's relationship count) but reads keep working. Only a failure
graded `heal: 'rebuild'` flips `serving` to `false` — **not-ready** — because the
provider itself is telling you its serving structure cannot answer correctly.
- **`invariants`** — each checked condition, with its provenance
(`source: 'ledger'` — an exact count; `'deep'` — a full scan, diagnostic-only;
`'unledgered'` — not yet tracked) and, for a failing one, an exact `missing`
count plus a capped sample of the affected ids — a verdict, never a dump.
- **`generation`** — bumps on every ledger mutation and rebuild, so a caller can
cache a verdict per generation instead of re-deriving it.
The distinction that matters day to day: `healthy: false` can be entirely benign —
a maintenance window, a divergence `repairIndex()` will clean up on its own
schedule. `serving: false` is not benign. It means this provider is refusing to
answer, on its own word, right now.
**How a failure gets its grade — the serving law.** A provider grades `heal` by
one question only: *could an answer be wrong?* — never *how expensive is the
fix?* A missing-postings shortfall, however large, is `heal: 'repair'` (re-post
exactly what the ledger names, reads serving throughout); it can never withhold
serving just because healing it takes work. `serving` is withheld only by a
small, named set of rebuild-graded conditions — the index not initialized, its
durable state absent, a manifest naming files that are not resident, a replay
that did not complete cleanly — the states in which an answer could genuinely be
wrong. And a read is only ever refused by the family it actually consults: a
metadata filter is answered by the metadata index alone, vector search by the
vector index, traversal by the graph index — one family's refusal never blocks
another family's reads.
## Reads refuse — they never rebuild
A query that reaches a not-serving provider does not trigger a rebuild from inside
the read. Brainy retired that path deliberately: a rebuild kicked off by an ordinary
`find({ where: { status: 'active' } })` call is a dark, unpredictable cost hiding
behind a request that looks like a cheap read. Instead, the read throws a typed,
catchable error naming the reason:
| Error | Thrown when | Meaning |
|---|---|---|
| `GraphIndexNotReadyError` | `find({ connected })`, `neighbors()`, `related()` | The graph adjacency index isn't serving — traversal would otherwise return `[]` indistinguishable from "no relationships" |
| `MetadataIndexNotReadyError` | `find({ where })` | The metadata/field index isn't serving — a filtered read would otherwise return `[]` indistinguishable from "no matches" |
| `VectorIndexNotReadyError` | `find({ query })`, `similar()` | The vector index isn't serving — a semantic search would otherwise return `[]` indistinguishable from "nothing similar" |
All three are exported from `@soulcraftlabs/brainy`. Catch them where your application
needs to distinguish "this index isn't ready yet" from "there's genuinely nothing
here" — a health dashboard, a retry policy, an operator alert. The fix is always
the same: reconcile the index, either by reopening the brain (which brings every
provider to serving before `init()` returns — see the next section) or by calling
description: How to back up a brain directory with external tools (tar, rsync, cp) without exploding sparse files — why a store can show 100+ GB "apparent" size on a small disk, which files are sparse, and how persist()/restore() handle it for you.
next:
- guides/snapshots-and-time-travel
- concepts/storage-adapters
---
# External Backups & Sparse Storage
The built-in snapshot path — [`db.persist()` and `brain.restore()`](/docs/guides/snapshots-and-time-travel) —
already handles everything on this page for you. Read this when you back up a brain directory with
**external tools**: `tar`, `rsync`, `cp`, `scp`, or a filesystem-level backup agent.
## The one-sentence rule
> **Always use the sparse-aware flag**: `tar czSf` (capital `S`), `rsync --sparse`,
> `cp --sparse=always`. A naive copy can turn a 2 GB store into a 100+ GB one — or fail
> the disk entirely.
## Why: some files are sparse
When a native accelerator plugin is active, parts of the index live in **memory-mapped files**
created at a large fixed virtual size — the file's *apparent* size — while the filesystem only
allocates blocks that were actually written. A brand-new id-mapper file can report tens of
gigabytes in `ls -l` while occupying a few megabytes on disk.
Check the difference yourself:
```bash
ls -lh brain-data/_id_mapper/ # APPARENT size (can be huge)
du -sh brain-data/ # ALLOCATED size (the real footprint)
@ -40,6 +40,10 @@ Brainy picks `maxLimit` from the first of these that's available:
Worked example: a 4 GB Cloud Run container picks priority 3 → `floor(4 GB × 0.25 / 25 KB) = floor(40 960) = 40 000` results. A 900 MB free-memory box on priority 4 gets `floor(900 MB / 25 KB) = ~36 000`.
The cap is fixed at construction and never changes at runtime. Query timing is recorded
for diagnostics only — a burst of slow queries cannot silently shrink the cap, and the
auto-detected tiers (3 and 4) never go below a floor of 10 000.
> **Calibration note.** Pre-7.30.2 used 100 KB per result instead of 25 KB, which produced caps that were 4× too tight for typical workloads (an 8 KB / result reality). 7.30.2 recalibrated to match observed entity sizes; existing `limit: 10_000` safety patterns now pass silently on any reasonably-sized box.
// For build-time usage (runs in Node during the build)
import { Brainy } from '@soulcraft/brainy'
import { Brainy } from '@soulcraftlabs/brainy'
export async function generateStaticProps() {
const brain = new Brainy({
@ -513,7 +513,7 @@ export async function generateStaticProps() {
### Issue: Large client bundle size
**Cause**: A client module is pulling in Brainy.
**Solution**: Move the `import { Brainy } from '@soulcraft/brainy'` into a server-only module so it never reaches the browser bundle.
**Solution**: Move the `import { Brainy } from '@soulcraftlabs/brainy'` into a server-only module so it never reaches the browser bundle.
### Issue: SSR hydration mismatch
**Solution**: Run the search on the server (loader / server action / API route) and pass the results down as props, so server and client render the same markup.
title: Migrating to 9.0 — your fields and system fields
slug: guides/namespace-migration
public: true
category: guides
template: guide
order: 1
description: The simple story of the 9.0 field-addressing change and the mechanical checklist for updating your call sites — every miss fails loudly with the fix in the error.
next:
- concepts/field-addressing
---
# Migrating to 9.0 — your fields and system fields
The one-sentence version: **your data's field names are now completely
yours, the engine's own fields all live behind one `system.` prefix, and
nothing in between can silently go wrong anymore.**
## What changed, simply
**1. Any field name just works.** Before 9.0 the engine quietly owned
certain names. A field called `level` could be shadowed by the engine's
internal index layer of the same name (sorts silently returned insertion
order); names like `confidence` or `subtype` were rejected inside
`metadata`; names like `content` or `id` were silently never indexed, so
filtering on them returned nothing. All of that is gone. Any name —
`level`, `confidence`, `type`, `id`, `content`, anything — is stored
exactly as written and works with every feature: filtering, sorting,
grouping, aggregation, search, and time-travel reads.
**2. The engine's fields moved behind `system.`.** The engine still keeps
its own per-record bookkeeping — creation time, type, confidence, and so
on. Those are reached one way only now: spelled out, e.g.
`system.createdAt`, `system.type`. They are just as queryable and sortable
as before. `orderBy: 'createdAt'` means *your* field named `createdAt`;
`orderBy: 'system.createdAt'` means the engine's timestamp. No guessing,
no priority rules.
**3. Storage keeps the two physically separate.** New records store your
metadata in its own nested compartment, so a user field named
`confidence` and the engine's confidence live side by side, both intact,
through restarts, index rebuilds, and `asOf()` history. Old records stay
readable forever; nothing rewrites your data.
**4. Mistakes are loud.** An ambiguous or unknown field name is a typed
error naming the fix. Unimplemented options refuse instead of being
ignored. The only forbidden name in your metadata is one literally
starting with `system.`.
## The mechanical checklist
Every missed site fails **loudly** with the correction in the error
message — nothing silently changes meaning. Sweep these patterns:
| `metadata: { confidence: 0.9 }` expecting a throw or a lift to the engine scalar | it is YOUR field now — set the engine scalar via the `confidence` param |
| `new Brainy({ reservedFieldPolicy: … })` | remove the option (it throws with this note) |
| `find({ cursor })` / `includeRelations` / `writeOnly` | refuse with `UnsupportedFindOptionError` — they were silently ignored before |
If a bare name in a query was genuinely *your* field all along (`orderBy:
@ -32,7 +32,7 @@ Brainy 7.31.0 adds a per-entity revision counter so multiple writers can coordin
Every distributed-job scheduler eventually wants this exact loop:
```ts
import { Brainy, RevisionConflictError } from '@soulcraft/brainy'
import { Brainy, RevisionConflictError } from '@soulcraftlabs/brainy'
const LOCK_ID = '...uuid for this job slot...'
@ -137,7 +137,7 @@ await brain.addIfMissing({ // ← not a real API
It's race-prone as a plain read-then-write: two concurrent imports both see "not found," both insert, you get duplicates. Without a unique-index primitive (which Brainy doesn't have today), close the race with whole-store CAS — read at a pinned generation, then commit only if nothing moved:
```ts
import { GenerationConflictError } from '@soulcraft/brainy'
import { GenerationConflictError } from '@soulcraftlabs/brainy'
async function addIfMissingByEmail(email: string, data: string) {
for (let attempt = 0; attempt <5;attempt++){
@ -180,3 +180,35 @@ Brainy 8.0 has exactly two write-coordination counters, at two granularities:
They compose: a `transact()` batch can carry per-entity `ifRev` checks *and* a whole-store `ifAtGeneration`; any failed check rejects the entire batch before anything is staged. Generations also power snapshots and time travel (`brain.now()`, `brain.asOf()`, `db.persist()`) — see the [consistency model](../concepts/consistency-model.md) and [Snapshots & Time Travel](./snapshots-and-time-travel.md).
A snapshot or historical view captures each entity *including* its `_rev` at that moment, so reading the past and writing back with `ifRev` against the live state works exactly as you'd hope: the write fails if the entity moved since the state you copied from.
## The transact envelope: batch size, budget, and bulk imports
`transact()` applies its batch atomically under one commit — which means the whole batch
shares one **apply budget**. Since 8.7.0 the budget scales with the batch:
`max(30 s, opCount × 2 s)`, or exactly what you pass as `timeoutMs`. A tripped budget rolls
the entire batch back (nothing partial survives) and throws a retryable
`TransactionTimeoutError` that names the operation it stopped at, the batch size, and the
elapsed vs budgeted time — a diagnosis, not just a failure:
```
Transaction timed out at operation 41/120 ('add') — 246012ms elapsed, budget 240000ms.
The batch rolled back atomically; retry with a higher timeoutMs or a smaller batch.
description: Subscribe to every committed mutation with `brain.onChange` — the in-process change feed behind live UIs, cache invalidation, and realtime sync. Covers the event shape, delivery guarantees, and catch-up patterns.
next:
- guides/optimistic-concurrency
- guides/snapshots-and-time-travel
---
# Reacting to Changes
`brain.onChange(cb)` is Brainy's in-process change feed: subscribe once and
receive one event per committed mutation — **every** mutation, regardless of
how it happened. Direct calls, batch methods, `transact()`, imports, and
Virtual Filesystem writes all funnel through the same commit point the feed is
and what's missing, stated) · 🔴 owed (named, never silent).
## LC — Lifecycle
| ID | Brainy row | Status |
|----|-----------|--------|
| LC1 | Same-version reopen adopts everything: brain-format epoch match → zero rebuilds; aggregation state adopts by stamp; persisted indexes load. | ✅ `tests/unit/brainy/brain-format-handshake` + `migration-deference` (no-drift reopen never rebuilds) |
| LC2 | New empty brain: doors immediate. | ✅ exercised by every suite's setup |
| LC3 | Upgrade, same epoch: as LC1 — new code on unchanged formats owes nothing at open. | ✅ same pins as LC1 (epoch equality is the gate) |
| LC4 | Upgrade with epoch migration: TODAY brainy's epoch rebuild runs at open before doors. | 🔴 **owed — the sev's lockout row.** The doors-open-serving-old-structures design (yielding installments + atomic swap) lands measured-and-gated behind the service-class pair, per the lifecycle-sprint choreography. Acceptance case: the 9,184-row hours-lockout. |
| LC5 | Crash recovery: bounded, resumable, narrated. Aggregation leg ✅ (behind-stamp → incremental catch-up off the fact log + time-travel reconciliation, capped at 5,000 affected before an ANNOUNCED rescan). Vector/metadata legs ride epoch machinery (rebuild-from-canonical, narrated). | 🟡 aggregation pinned (`tests/integration/aggregation-lifecycle-catchup`); the rebuild legs are narrated but not yet installment-yielding (couples to LC4) |
| LC6 | Shutdown under load: close() drains the background flush flight, tears down cadence timers, runs ONE time-bounded compaction pass (~5s budget, resumable). | 🟡 pinned for flush/compaction (8.9.0 suites); SIGTERM drain budget not yet declared |
| LC7 | Rollback/downgrade: an N−1 build opening an N brain. | 🔴 owed — no declared read-compat window or typed refusal today (epoch mismatch triggers a rebuild, not a refusal; v2 nested-bag records read as a phantom user field on pre-law builds). Needs the declared-window contract. |
| LC8 | Relocatable brain directory: no absolute paths in artifacts; persist()/load() round-trips. | 🟡 persist/load pinned; byte-for-byte relocation depot cases are the pair gate's (shared corpora) |
| LC9 | Double-open: second writer gets a typed lock refusal (PID-liveness + heartbeat stale detection; `force` escape hatch logs loudly). | ✅ writer-lock suites (8.7.1) |
## DP — Data plane
| ID | Brainy row | Status |
|----|-----------|--------|
| DP1 | `get()` by id: direct storage read + hydrate. INDEX-SERVED (id-mapped). Milliseconds at every scale. | ✅ exercised everywhere; budget rides the pair speed table |
| DP2 | `find({query})`: embed + vector search. The embed dominates (native side owns the budget); JS HNSW serves the search leg. | 🟡 300ms-class p95 is the pair speed-table row; brainy-alone budget declared there |
| DP3 | Filtered/sorted list: column top-K when the field is columnized (INDEX-SERVED, zero canonical reads on the sorted page — value pairs come from ONE batched metadata-record pass); no-column fallback is BOUNDED-ANNOUNCED (one batch pass, announces once per field past 500 rows); unknown field → TYPED REFUSAL naming both candidate spellings. | ✅ `tests/unit/utils/metadataIndex-sort-callshape` (zero per-row reads, batch-only — latency-blind) + `metadataIndex-nested-orderby` (dotted keys serve-or-refuse) + `tests/integration/orderby-sort-bug` |
| DP4 | Aggregation/stats: ALWAYS answers. Write-time incremental; behind-stamp reconciles incrementally; genuine rebuilds go through the native parallel door or the paged JS walk; nothing ever latches off; before-image-less deletes flag a LOUD rescan, never a silent skip. | ✅ `tests/integration/aggregation-lifecycle-catchup` + `tests/unit/aggregation/aggregation-provider-rebuild` |
| DP6 | Single write: ack at the canonical commit; visibility committed at ack (the atomic vector update kills the remove→add dark window); maintenance NEVER holds the ack (background flush cadence — THE ACK LAW pins: a hung flush cannot block a write, a hung EMBEDDER cannot block a write). | ✅ `tests/unit/brainy/persistence-policy` + `tests/unit/hnsw/update-item-atomic` + `tests/integration/deferred-embedding` |
| DP7 | Bulk ingest: sustained rate holds flat — per-write maintenance taxes must not grow with brain size (A4 removed caller-flush convoys; deferred embedding removes the per-write embed tax where opted). | 🟡 the decay-curve row is a pair speed-table RED GATE; brainy-alone sustained-rate run rides the same corpora |
| DP8 | Read under write pressure: no flicker window — a row that exists is never invisible to recall, even transiently (same-vector re-index is a no-op; changed-vector swaps in place, node never leaves the index; deferred updates serve the OLD vector until the atomic swap — stale-beats-absent). | ✅ brainy leg pinned (`tests/unit/hnsw/update-item-atomic` 9/9 + `deferred-embedding` stale-beats-absent); the symmetry property suite + runtime sentinels remain the B4 program |
| — | **As-of semantic recall** (time-travel vector search): `asOf(G).find()` serves the vectors AS THEY STOOD at G — byte-exact past vectors, tombstone masking, the deferred-embed cell honest on the vector leg, TYPED refusal beyond the head. Brainy-alone leg = ephemeral at-generation materialization (documented O(n log n at G) build, bounded); the at-scale leg rides the accelerated provider's as-of index. | ✅ `tests/integration/asof-semantic-recall` 4/4 (registry ID pending the master table's mint) |
| — | **The lazy-open gate honors EVERY provider's not-ready report** (a not-ready metadata provider can no longer latch the silent-empty state under `disableAutoRebuild`). | ✅ `tests/unit/brainy/lazy-notready-honor` |
## MT — Maintenance (never in the door path)
| ID | Brainy row | Status |
|----|-----------|--------|
| MT1 | Flush/checkpoint: ENGINE-OWNED cadence (write-count/interval/idle triggers, single-flight, background, loud on failure; callers never flush in hot paths; `flush()` stays as an awaitable barrier). | ✅ `tests/unit/brainy/persistence-policy` |
| MT3 | Index upkeep (mapper folds, delta promotion): native-side machinery; brainy's JS legs are small and synchronous-cheap. | 🟡 declared; yield audit rides the pair |
| MT4 | Heal/rebuild walks (`repairIndex`, backfill walks): paged; failure latches with cooldown; NOT yet yield-to-foreground installments. | 🔴 owed — the priority-isolation clause (couples to LC4; same choreography) |
| MT5 | Deferred embedding worker: ack at durability, durable pending markers (written BEFORE the commit — orphan-safe), crash-recovered at open via a bounded prefix listing, single-flight, 60s hang guard, `awaitPendingEmbeds()` barrier + `pendingEmbeds` gauge. VFS write paths adopt it end-to-end. | ✅ `tests/integration/deferred-embedding` 5/5 |
| MT6 | Retention/archival walks: retention `'all'` does nothing by design; bounded-retention reclaim is close-time/explicit only. | 🟡 8.9.0 behavior pinned; archival profile is the co-frozen D1+D3 unit |
## FM — Failure modes
| ID | Brainy row | Status |
|----|-----------|--------|
| FM1 | Disk full / IO error mid-op: transaction rollback + typed error; failed rollback → StoreInconsistentError quarantines writes until repairIndex(). | 🟡 rollback paths pinned; explicit disk-full depot case owed |
| FM3 | Torn/corrupt file on open: malformed brain-format marker → safe rebuild (never trusting a bad epoch); corrupt records surface loudly. | 🟡 marker pin ✅ (`brain-format-handshake`); broader quarantine is native-side |
| FM4 | Native module unavailable: plugin load failure is LOUD (version-coupling law throws on range mismatch — never silently version-drifted); JS engine serves with its own declared budgets, named as the active backend in op names. | ✅ `tests/unit/plugin-version-coupling` + op-name stamping |
## FL — Fleet
| ID | Brainy row | Status |
|----|-----------|--------|
| FL1 | Cold open on demand: LC1's adopt-everything open; warm() available for eager paths. | 🟡 open cost pinned at LC1; millisecond budget rides the speed table |
| FL2–FL4 | Boot storm / upgrade wave / isolation: fleet-layer policies over LC1/LC4 — engine leg = budgeted opens + LC4's behind-doors migration. | 🔴 owed with LC4 |
| `find` (semantic: embed + HNSW, 1k docs) | p50 178ms · p95 393ms | — | dominated by WASM query embedding (measured on a machine under concurrent load — treat the p95 as an upper bound); the vector search itself is single-digit ms |
## Write operations
Under Model-B **every write is its own durable generation** — a single-op `add` pays
serialization, before-image staging, and fsync before it acks. That durability is priced
| `open` (warm, populated, clean shutdown) | 763ms | 4.9s | pure-JS vector index load dominates and grows with entity count; the native accelerator exists precisely to remove this |
| `close` | bounded | bounded | auto-compaction pass is time-bounded (~5s max) since 8.9.0 |
A store that was NOT cleanly closed pays index rebuilds on top of the warm-open
number (tens of seconds at 10k) — clean shutdown is worth engineering for.
"description":"Universal Knowledge Protocol™ - World's first Triple Intelligence database unifying vector, graph, and document search in one API. Stage 3 CANONICAL: 42 nouns × 127 verbs covering 96-97% of all human knowledge.",
"headline":"Graph-first finds in production, and opens that stop rescanning history",
"items":[
"find({ connected, where }) now walks the neighbours first and filters only those rows through a native door — correct at every page and O(neighbours), never the whole store.",
"related() with a list of verb types returns every requested kind (a fast path had silently kept only the first).",
"Deferred-embedding recovery resumes from a low-water mark instead of rescanning the whole generation log at every open — measured at two minutes on a large brain, now milliseconds."
],
"url":null,
"thumb":null
},
{
"version":"11.0.4",
"date":"2026-09-01",
"headline":"Closes in milliseconds, index rebuilds without the disk-sync storm",
"items":[
"close() no longer pays deferred compaction or waits out an in-flight rebuild — measured 8 ms against the 4-minute closes it replaces; deferred work resumes at the next open, in the background.",
"The metadata index's rebuild syncs to disk per shard instead of per row, and the durability point moved to the publish step — the same guarantee, a fraction of the disk traffic.",
"A new native filter door evaluates queries over exactly the candidate rows a graph walk found, never the whole store."
],
"url":null,
"thumb":null
},
{
"version":"11.0.3",
"date":"2026-09-01",
"headline":"The embedding upgrade ceremony runs on every brain",
"items":[
"A brain opened through the standard plugin now carries its embedding-model identity, so the full-precision upgrade ceremony can run on it.",
"headline":"Hybrid finds filter before they hydrate, one owner per shutdown, and a faster open",
"items":[
"Hybrid finds (query/vector combined with a filter, including connected and fusion finds) now filter first and hydrate only the page — one batchGet of exactly the requested rows, instead of hydrating everything the search side found. Fixes a bug where any page after the first came back empty.",
"A brain now has exactly one shutdown owner — a host and its engine no longer race to close the same store, and a follow-up flush requested during a running flush is handed off cleanly instead of ever risking a stall.",
"find({ path }) and other path-scoped VFS searches now serve a real range over the indexed path (O(log n)) instead of refusing the query outright — both scoped and recursive:false searches were silently broken before this.",
"Open no longer rescans a brain's whole fact log on every open — sealed segments the manifest already accounts for are skipped, collapsing a multi-second open term to near-zero on large brains.",
"commitTransaction() now refuses by name if single-ops are still pending, and a read-only open no longer writes clean-shutdown evidence it didn't earn — two correctness invariants that were previously assumed, not enforced."
"headline":"A planner door for indexes, batched containment repair, and a fixed near()",
"items":[
"An optional planFindPage door lets an index plan a find() and answer it in one call, instead of the engine assembling the plan itself.",
"repairContainment's reconcile pass now walks paged edges once instead of issuing one graph call per file.",
"find({ near }) now searches around the anchor's own vector and refuses by name when none is available, instead of silently querying with no vector at all."
"headline":"Graph-first finds, honest verb arrays, and opens that stop rescanning history",
"items":[
"find({ connected, where }) now walks the neighbours first and filters only those rows — correct at every page, and O(neighbours) instead of O(store).",
"related() with a list of verb types (or sources, or targets) returns every requested kind — four fast paths silently kept only the first.",
"Deferred-embedding recovery resumes from a low-water mark instead of rescanning the whole generation log at every open — measured at two minutes on a large brain, now milliseconds."
'additive — a new optional door, a new served operator, a new error class; every existing implementation still conforms',
major:
'breaking — a door removed, an answer narrowed, an ordering law changed, an optional door promoted to required, or an operator moved from served to refused'
| `gate-preflight.sh` | machine is gate-clean | one or more `FATAL:` violations printed |
| `vitest-verdict-check.sh` | log's summary is trustworthy and matches | usage error, missing/unreadable log, or one or more `FATAL:` violations printed |
Non-zero from either script means: do not trust the gate that was about to
