One of the pain points when writing tests for Axum applications is debugging what went wrong. One thing I've been doing is keeping a test helper to setup a tracing subscribing on my test helpers to add to the test when debugging. i.e.

// t.rs
#[allow(unused)]
pub(crate) fn setup_tracing() {
    let _ = tracing_subscriber::fmt()
        .with_env_filter("debug,tower_http=trace")
        .with_test_writer()
        .try_init();
}
...
#[tokio::test]
async fn test_foo() -> Result<()> {
  t::setup_tracking();
  ...
}

I'm curious if people instead always setup tracing when running tests. And if so where would they enable it? AFAIU Rust's built-in test runner doesn't have something like Go's TestMain

I'm also curious what debugging strategies besides println! do y'all use.

1 post - 1 participant

Read full topic