This generally seems like the cleaner design because `SourceKitServer` is actually able to semantically inspect the message and decide whether it can be handled concurrently with other requests.
Since `ClangLanguageServerShim` calls directly into `SourceKitServer` we also no longer need the logic to forward a message from clangd to the editor in `handle`.
The core idea here is that the toolchain language servers always call into `BuildSystemManager` and `BuildSystemManager` will always reply with build settings. If it hasn’t computed them yet, it will reply with fallback settings.
With that assumption, we can remove the `documentToPendingQueue` from `SourceKitServer` since there are no longer any documents that are pending – everything has a build settings immediately.
Similarly, `BuildSystemManager.mainFileStatuses` also isn’t needed anymore.
And lastly, since we know that `BuildSystemManager.buildSettings` will always return a value `registerForChangeNotifications` is changed not call `fileBuildSettingsChanged` immediately. Instead, it will only cause `fileBuildSettingsChanged` to be called when the file’s build settings change after the `registerForChangeNotifications` call.
This defines away an entire class of data races if delegate callbacks are delivered out-of-order. If we aren’t providing the new build settings in the delegate callback, then it doesn’t matter if two `fileBuildSettingsChanged` calls change order since they don’t carry any state.
`LocalConnection` with its dynamic registration of a message handler made the overall design unnecessarily complicated. If we just call `SourceKitServer` from `ClangLanguageServerShim` and `SwiftLanguageServer` directly, it’s a lot more obvious, what’s going on, IMO.
Instead of storing build settings inside the language servers based on update notifications from the build system, always call into the `BuildSystemManager` to get the build settings.
Overall, I think this is a much clearer separation of concerns and will allow us to remove `SourceKitServer.documentToPendingQueue` in a follow-up commit as `SwiftLanguageServer` can always directly call into `BuildSystemManager` to get build settings and we don’t need to wait for the initial notification to receive the first build settings.
This requies `BuildServerBuildSystem` to keep track of the build settings it has received from the BSP server.
`ClangLanguageServer` still caches build settings locally. `ClangLanguageServer` will change to the same pull-based model in a follow-up commit.
Unfortuantely, we have a few potential out-of-order exeuction possibilities while we migrate everything else to also be asyncronous. But those should be low-probability issues that we can fix in follow-up commits, so I think it’s fine for now. All of these places are marked with `FIXME: (async)`
The asyncification changes caused some non-deterministic test failures. I believe that some of these are due to race conditions that are the result of the partial transition to actors.
Instead of merging the asyncification piece by piece, I will collect the changes asyncification changes in a branch and then qualify that branch througougly (running CI multiple times) before merging it into `main`.
Instead of storing build settings inside the language servers based on update notifications from the build system, always call into the `BuildSystemManager` to get the build settings.
Overall, I think this is a much clearer separation of concerns and will allow us to remove `SourceKitServer.documentToPendingQueue` in a follow-up commit as `SwiftLanguageServer` can always directly call into `BuildSystemManager` to get build settings and we don’t need to wait for the initial notification to receive the first build settings.
This requies `BuildServerBuildSystem` to keep track of the build settings it has received from the BSP server.
`ClangLanguageServer` still caches build settings locally. `ClangLanguageServer` will change to the same pull-based model in a follow-up commit.
