SQL logging during seeding

[jondot]

Discussed in #959

^{Originally posted by felipesere November 3, 2024}
Not sure if I'm holding it wrong 🤷‍♂️

I'm trying to move my PostsgreSQL backed app to use Sqlite (for "reasons") and my current hurdle is that I get the following error when seeding the users:

called `Result::unwrap()` on an `Err` value: DB(Json("premature end of input"))
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Is there a way to turn on SQL logging for seeding? I have it turned on in my test config but I see nothing.
It's hard to tell which column gives me grief :)

[felipesere]

I’m on my phone so can’t check this. But given that migrations are a separate crate with its own Cargo.toml, il pretty sure we would have to turn on the logging feature in the SeaORM dependency there…

I haven’t checked if it can be turned on at the crate level and then turned back off at runtime to be able to control it from the loco config.

[jondot]

no worries, i'll take a look as well

[jondot]

I did some checking, looks like logging and its flag wiring is good. Migration crate will get an already-initialized connection from the main Loco set up, so it will share what ever configuration were given initially.

I'm wondering if configuration on your end is really in sync with what you're running. For example, if you're running tests and expect to see these logs - your test.yaml might contain different flags, and also test.yaml has different log levels. I'd start with development.yaml for this.

The error itself follows from Sqlite requiring a Z to parse properly, indeed.

[jondot]

closing for now, please reopen if needed