Oneshot process exits before a sample can be captured

[gitstashpop]

I have a Rust program that makes a local request, gets a response, and exits. Using the local profiler, I was not able to see any jfrs, even with tuning the capture duration to 1ms.

A workaround I had was to put a sleep before exiting, and I was able to see jfr events.

Is it possible to always generate a report before exiting?

[arielb1]

Do you have a reproducer?

[arielb1]

When I run the example as the following:

LD_LIBRARY_PATH=path/to/libasyncProfiler cargo run --example simple -- --duration 200ms --local local-path

It works the sample when I use the S3 bucket uploader.

That calls the RunningProfiler::stop function, which performs the final JFR upload.

[rcoh]

Probably the cleanest option here is having the local reporter always report on drop if possible

[jlizen]

This should be addressed by #125, feel free to reopen if not. I'll cut a release shortly (#126)