Mattsson 5987523a25 fix(migrations): re-issue the 12xx label backfill inside an explicit transaction block (#2422) ...

* fix(migrations): re-issue the 12xx label backfill inside an explicit transaction block

20260908113353 (PR #2419) used a bare LOCK TABLE. The CI replay
(psql -f per file) and the Supabase branch runner execute migration
statements in autocommit, so Postgres refused it ("LOCK TABLE can only
be used in transaction blocks"): pg-real, pg-upgrade and tool-pg went
red on main and prod's migration queue stopped at that version, which
also blocks 20260908130127 (PR #2420). Prod never recorded 113353, so
the file is replaced rather than edited: same statements wrapped in
BEGIN/COMMIT (precedent 20260513140000), new version 20260908120449.
Applied and pg-tested on staging.

Refs #2413

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NXSuVejFCvRDyNXF1otEPd

* docs(decisions): migrations wrap transaction-only statements in BEGIN/COMMIT

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NXSuVejFCvRDyNXF1otEPd

---------

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>

2026-09-08 14:14:51 +02:00

..

pg

fix(migrations): re-issue the 12xx label backfill inside an explicit transaction block (#2422)

2026-09-08 14:14:51 +02:00

schema

fix(migration): complete-invoice-lines cron visits registers smallest first (#2341)

2026-09-06 18:38:31 +02:00

stubs

feat(analytics): add PostHog (EU) behind a same-origin proxy (#1237)

2026-07-27 14:30:49 +02:00

tool-pg

test(mcp): run the tools against a real PostgREST, not a fake supabase (#1983)

2026-08-27 18:11:41 +02:00

helpers.ts

feat(supplier-invoices): Inlagd i banken mark for payments entered by hand (#2220) (#2356)

2026-09-06 21:05:43 +02:00

pdf-text.ts

fix(pdf): keep the minus sign on losses in standard-font PDFs (#1982) (#1987)

2026-08-30 11:22:05 +02:00