David Barsky

> I have this test too but not sure how useful it would be > > ```rust > fn test_select_mode() { > // Test default behavior without the fbcode_build feature...

Ah, gotcah! Sorry for not responding, but I'm glad you were able to figure out the underlying issue.

> Please let me know if there's anything I can do to ease review on this. Also, is master the correct branch to target, or should I use v0.1.x if...

Oh, good to know. What is the error specifically?

> To make it clear, this is ready to be merged since there are breaking changes in master, it will just take some time for this to be actually realeased,...

I don't know why this is occurring right now, but [is this where](https://github.com/rust-lang/rust/blob/1f124e2ccf1d55b0062010f18a1efa062a0d8b20/compiler/rustc_log/src/lib.rs#L88) rustc sets up `tracing`? As an additional note, `tracing-core-0.1.31` did [introduce some](https://github.com/tokio-rs/tracing/releases/tag/tracing-core-0.1.31) changes that might be deadlock...

I'm going to close this; this is an issue in tracing-tree.

> I have a program where I update a specific field to reflect a state changing with time, and this approach would be a problem. In general, I do not...

I'm a little worried about this change: I'd _naively_ expect that the correct direction for "children outliving their parents" is `follows_from`, not changing how `on_exit` works.

> @davidbarsky how do you square that with the spec referenced above? Sorry for the delay. My stance is that, at least in this respect, `tracing` and the OpenTelemetry's specification...