* fix(sales-orders): pin the sales_order_items embed FK and teach the embed guard composite keys
Migration 20260902180000_sales_orders_hardening added a composite
(sales_order_id, company_id) foreign key from sales_order_items to
sales_orders next to the original single-column one. PostgREST then saw
two relationships and answered every `items:sales_order_items(*)` embed
with HTTP 300 / PGRST201, so kundorder list, detail, create and the MCP
list tool all failed on prod and staging with "Oväntat serverfel".
- Hint the three embeds with `!sales_order_items_sales_order_id_fkey`
(route, load service, MCP list tool).
- scripts/checks/ambiguous-embed.mjs only parsed single-column
`FOREIGN KEY (col)`, which is why the ratchet reported 0 for this pair.
It now reads composite column lists (named or default constraint
name) in both CREATE TABLE and ALTER TABLE, derives the same 17
ambiguous pairs prod's pg_constraint reports, and flags all three
shipped sites on main.
- Unit tests for the composite shapes: alongside a single-column key,
replacing one, and inline in CREATE TABLE.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Q7xJQL2aZo6iRHxCZNntUq
* fix(checks): drop composite embed edges when DROP COLUMN removes a member column
Postgres drops every foreign key a column takes part in, so the
ambiguous-embed parser must release a composite edge (and its constraint
name) when one of its columns is dropped, not only the single-column key.
Otherwise a later migration would keep a pair armed for a relationship
that no longer exists and reject valid embeds. Regression case added.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Q7xJQL2aZo6iRHxCZNntUq
---------
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Prod has three foreign keys between journal_entries and fiscal_periods, so
PostgREST answers PGRST201 to any embed of that pair that does not name the
relationship. pickAnchorEntry() destructured only data, so the error was
dropped and the helper returned null on every call since it shipped on
2026-07-27: supplier-invoice underlag has never once anchored in production.
Users see "Underlag saknas" on a verifikat that plainly shows the invoice PDF.
Names the constraint, matching the already-merged sibling fix in
lib/transactions/inbox-underlag.ts (6a40b3c0e), and handles the error instead
of dropping it.
Adds scripts/checks/ambiguous-embed.mjs to the ratchet guard, because neither
test layer can see this class: a mocked Supabase client never resolves a
relationship, and pg-real bypasses PostgREST entirely. The check derives the
ambiguous table pairs by parsing supabase/migrations, so a migration adding a
second foreign key between two tables arms the guard on the same commit; the
derived list reproduces prod's pg_constraint output exactly. It accepts both
PostgREST hint forms (constraint name and FK column name, both in use here) and
parses aliased embeds, which is the shape the real bug took.
Claude-Session: https://claude.ai/code/session_016ifKg6Ec67A39oxfGPU1yc
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Mattsson