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Outdated suggestions cannot be applied.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.
The server logs are plagued with the following error:
It appears that this is caused by
drain
being awaited by multiple coroutines. As a result this is non-trivial to reproduce on the test server as it requires a lot of messages to be sent at the same time. As far as I can tell this behavior is considered a bug in asyncio, as it is not mentioned in the documentation. A workaround, which is employed by thewebsockets
project, is to serialize the calls to drain with a mutex.I don't expect the addition of a lock to have a noticeable effect on performance as typically the calls to drain were intended to serve as a way of serializing messages per connection anyways (applying backpressure), the only place where this wouldn't naturally occur was in the broadcast functions, which are the reason for occasionally getting concurrent calls to drain.