andygrove opened a new pull request, #6300:
URL: https://github.com/apache/datafusion-comet/pull/6300
## Which issue does this PR close?
No issue. These are two doc corrections I found while auditing Comet's Tokio
usage, the audit that also filed #6292 and #6294.
## Rationale for this change
Two descriptions of the Tokio runtime no longer match the code. The
contributor guide says worker threads default to `num_cpus`, but since #3555
the count comes from `spark.executor.cores` or the thread count of a `local[N]`
master, and falls back to one outside local mode. The tuning guide already says
this.
`release_runtime`'s doc comment says the runtime is shut down in the
background so that the calling JNI thread is not blocked, but it calls
`shutdown_timeout(Duration::from_secs(3))`, which blocks the caller for up to 3
s. The comment also doesn't say what happens to work still running, which is
what #6294 is about.
## What changes are included in this PR?
- The worker-thread bullet in `development.md` now matches `tuning.md`.
- `release_runtime`'s doc comment now says:
- the shutdown blocks the calling thread for up to 3 s
- tasks still running are dropped at their next yield
- the next `init_runtime` or `get_runtime` call builds a new runtime
## How are these changes tested?
These are doc and comment changes only. `prettier --check` passes on the
edited page, and `rustfmt --check` passes on `jni_api.rs`.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
