Governed operator lists (organizations, grants, people, billing×4) gain
server-side search, status filters, and 50-row pages with true totals
from count(*) OVER(); state is URL-addressable, out-of-range pages
clamp,
and no-match is distinct from true-empty.
People is the eighth flat sidebar entry: /operator/persons lists persons
newest-joined first (excluding the reserved system person), rows linking
to the existing detail.
Billing gains an operator invoice detail at
/operator/billing/invoices/{invoiceID} reusing the member projection;
open invoices past due present as Overdue (derived, filterable, stored
status untouched); all four views lead with the linked organization and
mute object IDs.
Grants filter over the derived Live/Superseded/Inactive state, the SQL
HAVING predicate pinned to the Go derivation by test. Embedded lists
(org composite ledger, Tier changes) adopt the shared controls under
namespaced params with sibling-state-preserving URLs and scoped htmx
swaps that hold the viewport.
Review corrections: blocked ladder Delete renders disabled with tooltip
and mutations fire toasts; collapse triggers paint their open state;
sections use outside headings; plan topology drops the orphan-product
check; domains policy collapses behind a disclosure.
Replace domain claims with open invoices, show monthly recurring
revenue,
count team organizations, and add more informative trend captions.
Use standard bordered cards and add a README screenshot with reversible
sample-data tooling.
Back the operator landing surface with live deployment counts and
integration health signals, one sqlc query per owning module.
CountDeliveringGrants joins core.pool_provisions rather than filtering
grants.status: grants.status is an issuance ledger recording what was
written down and whether it was later revoked, not whether service is
flowing. The current-delivery fact lives on pool_provisions.status, and a
grant can sit at status='active' with every provision ended. Counting the
ledger alone would overstate delivery.
CountClaimsByLifecycle and CountOutboxByStatus each return their buckets
in a single row, so the halves that get printed together are read at the
same instant and the landing surface pays one round trip rather than one
per bucket.
Squash the pre-production migration history into fresh core, fedwiki,
and stripe baselines and reduce the canonical source list to those
three streams.
Update sqlc configs, generated queries, raw SQL, tests, and docs while
keeping provider tables schema-qualified.
BREAKING: existing local database volumes must be wiped because goose
version history restarts from the new baselines.
Update subscription model and SQL to include commitment fields. Add
SubscriptionScheduledChange types and queries, implement
UpdateSubscriptionItemPriceProduct, and adjust reconciliation and tests
to update core item rows when a Stripe item retains its ID but its
price/product/quantity changes.
Drop redundant schema prefixes from all sqlc-generated Go types. Since
each module generates into its own package, the package already provides
the namespace — billing.Account is unambiguous without
billing.BillingAccount.
Changes:
- Add rename: blocks to all 6 sqlc.yaml files mapping schema-prefixed
names to clean idiomatic names (e.g. BillingBillingAccount → Account,
IdentityPerson → Person, OrganizationOrganization → Organization)
- Rename billing.billing_accounts → billing.accounts (table name
repeated
the schema; the schema already provides that context)
- Rename integration.integration_outbox → integration.outbox (same
reason)
- Regenerate all sqlc output across billing, identity, organization,
entitlements, stripe, and fedwiki modules
- Update all calling code (server, workflows, provisioning, tests) to
use
the new names
- Add internal/db/sqlc_schemas.sql — sqlc-only schema declarations so
every module can resolve schema-qualified names without including the
full db migrations
- Update docs/database-management.md with the naming convention and
standard sqlc.yaml template
Convention going forward: table names must not repeat the schema name;
generated types carry no schema prefix; the Go package provides the
namespace (like http.Request, not http.HttpRequest).
cgalo5758