[tizzy]

I never gelled with how SQLC needs to know about your schema via the schema file. I'm used to flyway where you can update the schema as long as it's versioned correctly such that running all the sets of flyways will produce the same db schema.

I referred go-jet since it introspects the database for it's code generation instead.

[jchw]

The way I prefer to use sqlc is in combination with a schema migration framework like goose. It actually is able to read the migration files and infer the schema directly without needing an actual database. This seems to work well in production.

[bbkane]

That's how I'm using it as well (though I'm using some simple migration code instead of a framework): https://github.com/bbkane/enventory/tree/master/app/sqliteco...

I've been quite happy with this setup!