Fix: Rust Client Disconnect Race

[stevensJourney]

Overview

When disconnecting and connecting frequently, it is currently possible for the Rust sync clients to get locked in a frozen state if the StreamingSyncClient is disconnected/aborted before it has connected.

This can happen if:

• The client is triggered for a connect
• A disconnect is requested shortly after
• The StreamingSyncImplementation's AbortController is marked as aborted, but we nest the AbortController here.
• For WebSocket requests, we usually check if the provided AbortSignal has been aborted before making the request (and we throw), but due to the wrapped controller used in the Rust implementation we don't detect that the wrapped controller has been aborted. The wrapped controller will also never be aborted since the upstream controller won't emit the aborted event again (since it was aborted before)
• This causes the Rust client to remain connected. The ConnectionManager's disconnectingPromise never resolves, which causes a shared sync implementation to be frozen.

This checks the upstream AbortSignal before wrapping it. The abort is propagated by throwing a AbortOperation.

This could also have happened with the HTTP connection method. A check has been added there also.

[changeset-bot]

🦋 Changeset detected

Latest commit: 2f6153c

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 8 packages

Name	Type
@powersync/common	Patch
@powersync/adapter-sql-js	Patch
@powersync/node	Patch
@powersync/op-sqlite	Patch
@powersync/react-native	Patch
@powersync/tanstack-react-query	Patch
@powersync/web	Patch
@powersync/diagnostics-app	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[simolus3]

Nice catch, thanks!