Files
Behavision/server/migrations/013_references_are_immutable.sql
Suriyakumarvijayanayagam ce0223006b References are immutable, because clients now store them
012 turned three descriptive columns into identifiers other systems
keep: in agent.json on a shop counter, in a saved URL, in a scheduled
report. All three were already treated as stable and none of it was
enforced.

- clients.slug is an MQTT topic segment the broker ACL is written
  against. Rename one and that tenant's whole estate is silently refused
  by the broker, with no way to tell the agents.
- sites.slug is what a shop PC calls itself - agent.json holds
  "site_id": "chennai", never the uuid. A rename orphans the PC from the
  shop it is standing in.
- site_cameras.camera_id lands in visits.camera_id, which is text and
  not a foreign key. A rename orphans every visit already attributed to
  the old name: the footfall is still there and no longer joins to a
  camera. This was half-enforced in handleUpdateCamera and nowhere else,
  which is the shape of a rule that holds until somebody adds a second
  write path.
- visitors.number is assigned once from the tenant's counter and read
  back as V-42.

A trigger, not a CHECK: a CHECK cannot see the old row and the rule is
about the transition. The DISPLAY name is deliberately not frozen -
"TeNext Chennai", "Front door" - it is what a person reads, nothing keys
on it, and a system that cannot fix a typo in a shop's name has confused
the two.

Also records why the uuid stays where a slug would do. The length was
never the problem; needing it was, and that is fixed. Replacing it would
touch eight foreign keys on a live database to shorten a field clients
are already told not to use, and a sequential id would make any future
tenancy hole walkable by counting. It is NOT because ids must be minted
offline - sites, visitors and visits are all created server-side with a
database in hand, and claiming otherwise would defend the status quo
rather than explain it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HViLj9gYNRtSr7YVZmW5sn
2026-09-07 12:17:49 +05:30

98 lines
4.2 KiB
PL/PgSQL

-- The references clients are now told to use must not be able to change.
--
-- 012 made `chennai`, `Office1` and `V-42` first-class: every route that takes
-- an id takes one of these instead, and the documentation tells a client to
-- prefer them. That turns three columns which were merely descriptive into
-- IDENTIFIERS other people store - in an agent's config file on a shop counter,
-- in a saved URL, in a report somebody scheduled.
--
-- All three were already treated as stable, and none of it was enforced:
--
-- * `clients.slug` appears in MQTT topics as bv/<client>.<site>/... and the
-- broker ACL is written against it. Renaming one silently stops that
-- tenant's estate from being able to publish, and the agents cannot be told
-- - they would simply be refused by the broker.
-- * `sites.slug` is what a shop PC calls itself: agent.json holds
-- "site_id": "chennai". A rename orphans the PC from the shop it is
-- standing in.
-- * `site_cameras.camera_id` is what lands in `visits.camera_id`, which is a
-- text column and not a foreign key. Renaming it orphans every visit
-- already attributed to the old name - the footfall is still there and no
-- longer joins to a camera.
--
-- The camera case was half-enforced in one handler (`handleUpdateCamera` nils
-- CameraID before saving) and nowhere else, which is the shape of a rule that
-- holds until somebody adds a second write path. This is the backstop, in the
-- one place every write has to go through.
--
-- Deliberately a trigger and not a CHECK: a CHECK cannot see the old row, and
-- the rule is about the transition, not the value.
--
-- Note what this does NOT freeze. `name` - "TeNext Chennai", "Front door" - is
-- free to change and always should be: it is what a person reads, it is not what
-- anything keys on, and conflating the two is how a system ends up unable to fix
-- a typo in a shop's name.
BEGIN;
CREATE OR REPLACE FUNCTION reference_is_immutable() RETURNS trigger
LANGUAGE plpgsql AS $$
BEGIN
RAISE EXCEPTION
'% is a public reference and cannot be changed (% -> %); '
'create a new row instead, or change the display name',
TG_ARGV[0], OLD.slug, NEW.slug
USING ERRCODE = 'check_violation';
END;
$$;
-- camera_id lives in its own function only because the column is named
-- differently; splitting it keeps the message honest about which value moved.
CREATE OR REPLACE FUNCTION camera_reference_is_immutable() RETURNS trigger
LANGUAGE plpgsql AS $$
BEGIN
RAISE EXCEPTION
'camera_id is a public reference and cannot be changed (% -> %); '
'every visit already recorded names the old one. Change the label instead',
OLD.camera_id, NEW.camera_id
USING ERRCODE = 'check_violation';
END;
$$;
DROP TRIGGER IF EXISTS clients_slug_immutable ON clients;
CREATE TRIGGER clients_slug_immutable
BEFORE UPDATE OF slug ON clients
FOR EACH ROW WHEN (OLD.slug IS DISTINCT FROM NEW.slug)
EXECUTE FUNCTION reference_is_immutable('clients.slug');
DROP TRIGGER IF EXISTS sites_slug_immutable ON sites;
CREATE TRIGGER sites_slug_immutable
BEFORE UPDATE OF slug ON sites
FOR EACH ROW WHEN (OLD.slug IS DISTINCT FROM NEW.slug)
EXECUTE FUNCTION reference_is_immutable('sites.slug');
DROP TRIGGER IF EXISTS site_cameras_id_immutable ON site_cameras;
CREATE TRIGGER site_cameras_id_immutable
BEFORE UPDATE OF camera_id ON site_cameras
FOR EACH ROW WHEN (OLD.camera_id IS DISTINCT FROM NEW.camera_id)
EXECUTE FUNCTION camera_reference_is_immutable();
-- A visitor's number is assigned once from the tenant's counter and read back
-- as V-42. Nothing writes it after the insert; this says so.
DROP TRIGGER IF EXISTS visitors_number_immutable ON visitors;
CREATE OR REPLACE FUNCTION visitor_number_is_immutable() RETURNS trigger
LANGUAGE plpgsql AS $$
BEGIN
RAISE EXCEPTION
'visitors.number is a public reference and cannot be changed (% -> %)',
OLD.number, NEW.number
USING ERRCODE = 'check_violation';
END;
$$;
CREATE TRIGGER visitors_number_immutable
BEFORE UPDATE OF number ON visitors
FOR EACH ROW WHEN (OLD.number IS DISTINCT FROM NEW.number)
EXECUTE FUNCTION visitor_number_is_immutable();
COMMIT;