migration.record_invalid

A file in a migration's folder is missing, can't be read, or was changed by hand, so Jadpo can't use the migration.

Area 5, Entities, stores and migrations; not declinable; the next step is usually the human's.

Why Jadpo refuses it

jadpo migrate new writes each migration's folder: snapshot.json, the whole schema at that point, and postgres.sql, the SQL that gets there. jadpo check compares the entities with the latest snapshot, and the next migration is worked out from it. A record that can't be read breaks both, and a guess at what it held could change a real database in a way nobody wrote.

Jadpo won't rebuild the file for you. The right content is the one that was committed, and that may already have been applied to a database.

What it looks like

A project whose first migration's snapshot was edited by hand:

database/migrations/20261020T1400_initial/snapshot.json:

{"format": 1}

(human) at the end of the next: line tells your coding agent to stop and ask you which fix you want, because the choice is yours to make: it decides which committed version of the migration your databases are compared with.

How to fix it

If the file was committed, bring back the committed version:

git restore database/migrations/20261020T1400_initial/snapshot.json

The snapshot again matches what jadpo migrate new wrote, so jadpo check can compare the entities with it, and the next migration starts from it.

A migration's files are never fixed by hand, since jadpo migrate new is the only thing that writes them.

Related

• Migrations
• Migrations, in Review and records