util: ensure wait_for can only be used in a non-async runtime #61
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jayshrivastava
force-pushed
the
js/ntran/tests/copy
branch
2 times, most recently
from
545c4f4 to
269516f
Compare
jayshrivastava
changed the title
Previously, `wait_for` would deadlock when used in an async runtime (ie. any async function calls `wait_for` directly or indirectly). Now, `wait_for` returns an error if called in an async runtime. There's no reason to call it in an async runtime / function because you can simply await the future. This change updates a few places where `wait_for` is called to be async functions and just wait on the future. Why did the old `Spawner` deadlock? `Handle::current().block_on(f)` and the `std::sync::mpsc::channel` would both block an executor thread in tokio (this is a very bad practice generally). If there's no executer threads available, you can't run anything. Now, the closure just does `f.await` instead of `block_on`, which does not block an executor. The channel is now a `tokio::sync::mpsc::channel` as well with a buffer size of 1, so the async send will not block (I think a buffered std channel would have worked, but it's better to use tokio implementations generally). The receiver of the channel is in a sync runtime, so it does a blocking receive. Testing - unit tests now pass - added a unit test for `wait_for` returning an error in an async runtime - test for nested `wait_for` calls which now return an error Closes https://github.com/datafusion-contrib/datafusion-distributed/issues/63
jayshrivastava
force-pushed
the
js/ntran/tests/copy
branch
from
269516f to
335f975
Compare
jayshrivastava
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previously,
wait_forwould deadlock when used in an async runtime (ie. any async function callswait_fordirectly or indirectly). Now,wait_forreturns an error if called in an async runtime. There's no reason to call it in an async runtime / function because you can simply await the future. This change updates a few places wherewait_foris called to be async functions and just wait on the future.Why did the old
Spawnerdeadlock?Handle::current().block_on(f)and thestd::sync::mpsc::channelwould both block an executor thread in tokio (this is a very bad practice generally). If there's no executer threads available, you can't run anything. Now, the closure just doesf.awaitinstead ofblock_on, which does not block an executor. The channel is now atokio::sync::mpsc::channelas well with a buffer size of 1, so the async send will not block (I think a buffered std channel would have worked, but it's better to use tokio implementations generally). The receiver of the channel is in a sync runtime, so it does a blocking receive.Testing
wait_forreturning an error in an async runtimewait_forcalls which now return an errorCloses https://github.com/datafusion-contrib/datafusion-distributed/issues/63