Mark Van de Vyver

Apologies, @andylokandy I overlooked this: > If the only use case is to spawn a new span... The use case is to ***avoid*** spawning a new span. Rather, it returns...

Without knowing the impl details I assume the performance impact/cost of implementing `Span::current()` will be the same as the Tracing crate. One idea that might eliminate any performance impact is...

My reservation with `minitrace::event()` is related to OpenTelemetry. My understanding is that the `add_property()` ( or `record` in tracing) stashes data in Tags (or what show up as Tags in...

> I am not sure whether introducing a local variable to the generated span is a good design. Ack the concern, that did make me think a little harder, and...

@andylokandy not sure I understand your thread workaround example [here](https://github.com/tikv/minitrace-rust/issues/117#issuecomment-1070913840). As best I can tell it is not possible to move the `thread::spawn()` inside `do_stuff()`. Correct?

This will likely help keep embedded use cases as first class citizens? See issue #131 We're not in that space, but still do appreciate the performance improvements they squeeze out...

I also see this behavior in the VS-Code IDE. Not sure what the plans are for the `collect:` parameter.... if it is intended to allow this to be passed-in, in...

Collating functions similarly affected: - [ ] `root` - [ ] [`enter_with_local_parent`](https://github.com/tikv/minitrace-rust/blob/52891cc0deef525601220ff1d55649d40542ae63/minitrace/src/span.rs#L90)

High priority IMO. Regressions are devilishly difficult to identify after the fact. The criterion author had a crate targeting CI use cases. Let me try to dig it up.

IMO the approach to take is to create benchmarks using [iai](https://github.com/bheisler/iai) rather than criterion - not that you can't do both. However, I believe if you want to accept/reject/evaluate the...