The broadcast channel does not require values to implement `Sync` yet it calls
the `.clone()` method without synchronizing. This is unsound logic. This patch
adds per-value synchronization on receive to handle this case. It is unlikely
any usage of the broadcast channel is currently at risk of the unsoundeness
issue as it requires accessing a `!Sync` type during `.clone()`, which would be
very unusual when using the broadcast channel.
See #5595 and #7172.
This adds a debug assertion that checks that a supplied underlying std socket is set to nonblocking mode when constructing a tokio socket object from such an object.
This only works on unix.
Promote the feature flags and supported platforms sections out from
under examples, as they are not examples. Adjust their subsections
accordingly. Expose these subsections via navigation sidebar.
* make self-tracing public
* address review comments
* try to fix doctest
* adjust imports to fit standard
* more documentation
---------
Co-authored-by: Ariel Ben-Yehuda <arielby@amazon.com>
When a `?` operator is used in a tokio entry point function (wrapped in
`#[tokio::main]`), which has a Option or Result return type, but where
the function does not actually return that type correctly, currently the
compiler returns two errors instead of just one. The first of which is
incorrect and only exists due to the macro expanding to an async block.
```
cannot use the `?` operator in an async block that returns `()`
```
This commit is a characterization test for this behavior to help show
when it's fixed (or even changed for better / worse)
The test's previous iteration could sometimes flake since we didn't
await the completion of the first task. Since the tasks only existed to
`move` the relevant `rx`'s in, to force a drop, we can omit them
entirely and drop the `rx`s via `drop()`. This prevents any
scheduling-related flakes.
