Removes operator-private material from the shipped tree, corrects documentation
that had drifted from the code, deletes dead code, and applies a consistent
prose style.
Private material moved out of the public tree (docs/private/, git-ignored):
- the wallet keyfile recovery runbook and its helper scripts, which together
published the mainnet wallet file path, the service account, backup snapshot
filenames, and the wallet RPC listener state. The helper duplicated the same
constants, so relocating the document alone would not have removed them.
- the stagenet parallel-candidate runbook and its fixtures.
- the dated production-state record, the internal remediation plan, and the
TODO list. None are reference documentation for an outside reader.
Redacted or genericised:
- operator home-directory paths in shipped scripts and units. The asset build
script now takes KENNEY_ASSETS_DIR instead of a hard-coded path.
- private LAN addresses in tests, env examples and deploy docs, replaced with
RFC 5737 documentation addresses.
- the default alert sender and the notification hub example.
- the Monero wallet unit now matches the Wownero one: a dedicated service
account and a config file, instead of the operator's login account with the
wallet password in argv where any local user can read it from /proc.
Ignore rules now cover every .env variant rather than the bare name, plus local
agent state. The deployment docs instruct operators to create .env.mainnet and
.env.stagenet, which the exact-name rule would have allowed to be committed.
Documentation corrected against the code: per-chain dungeon level counts, the
premium unlock rule, Tavern chat scope, backup retention, the chain-profile
fallback, the environment template location, and several sections describing
shipped work as pending.
Dead code removed: configPersistence, rpccalls, a duplicate vendored ROT.js
bundle, the unused tile-picker overlay, the unreachable legacy game-size path
and its vestigial parameter, spectator stubs that shadowed the real
SpectatorManager, and commented-out debugging. nip98 now imports the shared
nostr loader rather than duplicating it.
Prose style: em-dashes removed throughout in favour of colons, semicolons and
commas; comments restated as present-tense descriptions of the code rather than
narration of past changes.
127 suites / 1015 tests pass.
4.2 KiB
Historical financial constraint release gate
Several migrations add CHECK ... NOT VALID. PostgreSQL enforces those constraints for every new and
changed row, but it does not scan rows that already existed when the constraint was added. A complete
schema_migrations ledger therefore does not, by itself, prove that historical financial rows obey
the current constraints.
Two repository-owned gates close that evidence gap without changing a live database. Both cover the
same 18 constraints across payments, payouts, games, matches, match_queue_entries,
race_entry_lots, and payment_refunds.
1. Read-only historical audit
scripts/deploy/financial-constraint-audit.sql checks the exact connected database name, verifies
the complete 18-constraint catalog inventory, and reports a named row count for every historical
predicate. It runs in one repeatable-read, read-only transaction, takes no explicit table lock, and
rolls back. Connection and credential handling stay operator-specific; keep database passwords out of
command history.
psql -X -v ON_ERROR_STOP=1 \
--set=expected_database=EXACT_DATABASE_NAME \
--file=scripts/deploy/financial-constraint-audit.sql
Exit codes:
| Code | Meaning |
|---|---|
0 |
Every required constraint exists and every historical row satisfies it. |
2 |
expected_database was not supplied, or the connected database does not match it exactly. |
3 |
A required constraint is missing, or at least one predicate has a nonzero violation count. |
The violation predicates mirror PostgreSQL CHECK semantics: only FALSE counts as a violation,
since CHECK accepts NULL. Reconcile any violation with a separately reviewed, backup-backed
repair. The audit has no repair mode by design.
This check is safe against production and should also run against every restored release candidate database. It reports how many of the constraints are still catalog-marked unvalidated; that count is informational, because the second gate exercises PostgreSQL's own validator.
2. Explicit validation on a disposable restore
scripts/deploy/financial-constraint-validate-restore.sql includes the read-only audit, then issues
an explicit ALTER TABLE ... VALIDATE CONSTRAINT for all 18 constraints inside a transaction. It
confirms pg_constraint.convalidated for each one and rolls the whole transaction back, so no
catalog flag changes durably.
VALIDATE CONSTRAINT still scans tables, takes locks, and transiently updates PostgreSQL catalogs.
The script therefore refuses to reach BEGIN unless all three conditions hold:
- the connected database name exactly equals
expected_database; - that name contains
restoreorscratch; - the exact disposable-restore confirmation string is supplied.
Run it only against a freshly restored, isolated database with no application traffic:
psql -X -v ON_ERROR_STOP=1 \
--set=expected_database=monerogue_release_restore \
--set=confirm_disposable=I_CONFIRM_THIS_DATABASE_IS_A_DISPOSABLE_RESTORE \
--file=scripts/deploy/financial-constraint-validate-restore.sql
Both scripts set statement_timeout to 10 minutes and lock_timeout to 5 seconds for the duration
of their transaction.
A pass proves both the explicit historical predicates and PostgreSQL's native validation path on that
restored data. It does not mark live constraints validated. Making a live VALIDATE CONSTRAINT
durable is a separate operation that needs its own reviewed maintenance plan, backup and restore
proof, lock budget, and rollback decision. Do not fold that mutation into application startup.
Static coverage
A focused test inventories every migration occurrence of NOT VALID and fails if either gate omits a
constraint or if the inventory size drifts from 18:
npm --prefix src test -- --runTestsByPath ../test/financialConstraintValidation.test.js
The test also asserts that the audit script contains no ALTER TABLE and no write statement, so the
read-only gate cannot silently gain a mutation.
Adding another NOT VALID constraint requires three matching updates: its exact historical predicate
and inventory entry in the audit script, its VALIDATE CONSTRAINT statement and inventory entry in
the restore script, and its table and name in the test's constraint list.